From 5de50b155ac90653e72a4319b22fdf2a7115f01f Mon Sep 17 00:00:00 2001
From: meels <meelstorm@gmail.com>
Date: Tue, 28 Jul 2026 15:20:55 +0200
Subject: [PATCH] chore: renormalize line endings under the new .gitattributes

Line-ending only. Verified: 3667 insertions against 3667 deletions with
zero content difference under --ignore-cr-at-eol.
---
 .obsidian/plugins/webinar-dash/main.js        | 1202 +++---
 .obsidian/plugins/webinar-dash/manifest.json  |   18 +-
 .obsidian/plugins/webinar-dash/styles.css     |  478 +--
 .../webinar-dash/test/coverage.test.js        |  244 +-
 .../webinar-dash/test/pipeline.test.js        |  218 +-
 .../plugins/webinar-dash/test/safety.test.js  |   94 +-
 CLAUDE.md                                     |  496 +--
 .../plans/2026-07-28-webinar-dashboard.md     | 3490 ++++++++---------
 .../2026-07-28-webinar-dashboard-design.md    |  512 +--
 index.md                                      |  178 +-
 log.md                                        |  314 +-
 wiki/script-coverage.md                       |   90 +-
 12 files changed, 3667 insertions(+), 3667 deletions(-)

diff --git a/.obsidian/plugins/webinar-dash/main.js b/.obsidian/plugins/webinar-dash/main.js
index d5f5464..a64dbc6 100644
--- a/.obsidian/plugins/webinar-dash/main.js
+++ b/.obsidian/plugins/webinar-dash/main.js
@@ -1,601 +1,601 @@
-"use strict";
-
-// Obsidian injects its own module resolver. Under plain `node --test` it is
-// absent, so guard the require and fall back to an empty base class. This is
-// what keeps the pure helpers below unit-testable without an Obsidian runtime.
-let OB = null;
-try {
-  OB = require("obsidian");
-} catch (_) {
-  OB = null;
-}
-const PluginBase = OB ? OB.Plugin : class {};
-
-const STATIONS = ["Chat box", "ReAct", "Tools", "Memory", "Skills", "Process", "OS"];
-
-const DEFAULTS = {
-  script: "raw/sources/Webinar script.md",
-  coverage: "wiki/script-coverage.md",
-  rawDir: "raw/sources",
-  wikiSourceDir: "wiki/sources",
-  conceptDir: "wiki/concepts",
-};
-
-const RAW_PATH_RE = /^\s*-\s*\*\*Raw path:\*\*\s*`([^`]+)`/m;
-
-function extractRawPath(text) {
-  const m = String(text).match(RAW_PATH_RE);
-  return m ? m[1].trim() : null;
-}
-
-function derivePipeline({ rawFiles, sourcePages }) {
-  // First claim on a raw path wins. A later page claiming the same file is a
-  // duplicate claim — real catalog drift — and joins `orphaned` rather than
-  // being silently dropped. `orphaned` therefore means "source page not paired
-  // with a raw file", whatever the reason.
-  const claimed = new Map();
-  const duplicates = [];
-  for (const page of sourcePages) {
-    if (!page.rawPath) continue;
-    if (claimed.has(page.rawPath)) duplicates.push(page);
-    else claimed.set(page.rawPath, page);
-  }
-
-  const processed = [];
-  const unprocessed = [];
-  for (const file of rawFiles) {
-    const page = claimed.get(file.path);
-    if (page) processed.push(Object.assign({}, file, { page }));
-    else unprocessed.push(file);
-  }
-
-  // One pass over sourcePages, so a page appears in `orphaned` at most once no
-  // matter how many of the three reasons apply to it. Concatenating a separate
-  // duplicates array here would double-count a losing claimant whose shared raw
-  // path is also missing from disk.
-  const rawPaths = new Set(rawFiles.map((f) => f.path));
-  const duplicateSet = new Set(duplicates);
-  const orphaned = sourcePages.filter(
-    (p) => duplicateSet.has(p) || !p.rawPath || !rawPaths.has(p.rawPath)
-  );
-
-  unprocessed.sort((a, b) => a.name.localeCompare(b.name));
-  processed.sort((a, b) => b.page.name.localeCompare(a.page.name));
-
-  return { processed, unprocessed, orphaned };
-}
-
-const VALID_STATUS = new Set(["covered", "partial", "absent"]);
-const SEPARATOR_RE = /^\|?\s*:?-{2,}/;
-
-function parseCoverageTable(markdown) {
-  const text = String(markdown);
-  const meta = { script: null, lastSynced: null };
-
-  const scriptM = text.match(/^\s*-\s*\*\*Script:\*\*\s*`([^`]+)`/m);
-  if (scriptM) meta.script = scriptM[1].trim();
-  const syncM = text.match(/^\s*-\s*\*\*Last synced:\*\*\s*(\S+)/m);
-  if (syncM) meta.lastSynced = syncM[1].trim();
-
-  const rows = [];
-  const errors = [];
-  let inTable = false;
-
-  text.split(/\r?\n/).forEach((line, i) => {
-    const t = line.trim();
-    if (!t.startsWith("|")) {
-      inTable = false;
-      return;
-    }
-    if (SEPARATOR_RE.test(t)) {
-      inTable = true;
-      return;
-    }
-    if (!inTable) return;
-
-    const body = t.endsWith("|") ? t.slice(1, -1) : t.slice(1);
-    // Split on unescaped pipes only. Obsidian escapes the pipe of a piped
-    // wikilink inside a table cell as `\|`, and a plain split("|") tears
-    // `[[harness\|alias]]` into two cells, shifting every later column left.
-    const cells = body.split(/(?<!\\)\|/).map((c) => c.trim());
-
-    if (cells.length < 3) {
-      errors.push({ line: i + 1, text: t, reason: "expected at least 3 columns" });
-      return;
-    }
-
-    const status = cells[1].toLowerCase();
-    if (!VALID_STATUS.has(status)) {
-      errors.push({ line: i + 1, text: t, reason: `invalid status "${cells[1]}"` });
-      return;
-    }
-
-    // Capture up to the first ] | or backslash. Obsidian escapes the pipe in a
-    // piped wikilink inside a table cell, so the raw cell reads [[name\|alias]] —
-    // excluding the backslash is what keeps the trailing "\" out of the name.
-    const linkM = cells[0].match(/\[\[([^\]|\\]+)/);
-    const stationCell = cells[2];
-
-    rows.push({
-      concept: linkM ? linkM[1].trim() : cells[0],
-      status,
-      stations:
-        stationCell === "—" || stationCell === "-" || stationCell === ""
-          ? []
-          : stationCell.split(",").map((s) => s.trim()).filter(Boolean),
-      pinned: (cells[3] || "").toLowerCase() === "yes",
-      line: i + 1,
-    });
-  });
-
-  return { meta, rows, errors };
-}
-
-function groupByStation(rows) {
-  const order = ["All stations", ...STATIONS, "No station"];
-  const buckets = new Map(order.map((k) => [k, []]));
-
-  for (const row of rows) {
-    let key;
-    if (row.stations.includes("all")) key = "All stations";
-    else if (row.stations.length === 0) key = "No station";
-    else key = row.stations[0];
-    if (!buckets.has(key)) buckets.set(key, []);
-    buckets.get(key).push(row);
-  }
-
-  const known = order.filter((k) => buckets.get(k).length > 0);
-  const unknown = [...buckets.keys()].filter((k) => !order.includes(k) && buckets.get(k).length > 0);
-  return [...known, ...unknown].map((station) => ({ station, rows: buckets.get(station) }));
-}
-
-function reconcileConcepts(rows, conceptNames) {
-  const named = new Set(conceptNames);
-  const rowed = new Set(rows.map((r) => r.concept));
-  return {
-    rows,
-    unsynced: conceptNames.filter((n) => !rowed.has(n)).sort(),
-    stale: rows.filter((r) => !named.has(r.concept)),
-  };
-}
-
-// Defence in depth. The spawn path no longer uses a shell, so nothing here is
-// load-bearing against injection today — but the guard costs nothing and would
-// still hold if the single-string `shell: true` form is ever needed for a .cmd
-// shim. `^` is cmd.exe's escape character; control characters are rejected
-// because a NUL byte makes spawn() throw synchronously.
-//
-// Apostrophes, spaces, cyrillic, em dashes and `!` stay allowed — they appear in
-// real filenames in this vault.
-const UNSAFE_CHARS = /["`$&|;<>%^\u0000-\u001f\u007f]/;
-
-function isSafeFilename(name) {
-  if (typeof name !== "string" || name.length === 0) return false;
-  if (UNSAFE_CHARS.test(name)) return false;
-  if (name.includes("..")) return false;
-  return true;
-}
-
-function parseConfig(source) {
-  const cfg = Object.assign({}, DEFAULTS);
-  for (const line of String(source).split(/\r?\n/)) {
-    const m = line.match(/^\s*([A-Za-z][A-Za-z0-9_]*)\s*:\s*(.+?)\s*$/);
-    if (m && Object.prototype.hasOwnProperty.call(DEFAULTS, m[1])) {
-      cfg[m[1]] = m[2];
-    }
-  }
-  return cfg;
-}
-
-// Source types CLAUDE.md documents for raw/sources. Anything else in that
-// folder is not a source and stays out of the queue.
-const RAW_EXTENSIONS = new Set(["md", "txt", "pdf"]);
-
-const ICONS = {
-  terminal: "M4 17l6-6-6-6M12 19h8",
-  check: "M20 6 9 17l-5-5",
-  minus: "M5 12h14",
-  x: "M18 6 6 18M6 6l12 12",
-  refresh: "M3 12a9 9 0 0 1 9-9 9 9 0 0 1 6.7 3L21 8M21 3v5h-5 M21 12a9 9 0 0 1-9 9 9 9 0 0 1-6.7-3L3 16M3 21v-5h5",
-};
-
-// Built with createElementNS rather than Obsidian's createSvg helper, whose
-// availability varies by version. This works on any Obsidian build.
-const SVG_NS = "http://www.w3.org/2000/svg";
-
-function addIcon(parent, name) {
-  const svg = document.createElementNS(SVG_NS, "svg");
-  svg.setAttribute("viewBox", "0 0 24 24");
-  svg.setAttribute("fill", "none");
-  svg.setAttribute("aria-hidden", "true");
-  const path = document.createElementNS(SVG_NS, "path");
-  path.setAttribute("d", ICONS[name]);
-  path.setAttribute("stroke", "currentColor");
-  path.setAttribute("stroke-width", "1.75");
-  path.setAttribute("stroke-linecap", "round");
-  path.setAttribute("stroke-linejoin", "round");
-  svg.appendChild(path);
-  parent.appendChild(svg);
-  return svg;
-}
-
-function formatBytes(n) {
-  return `${n.toLocaleString("en-US")} B`;
-}
-
-function renderLeftPane(container, pipeline, onIngest, gate) {
-  const pane = container.createDiv({ cls: "wd-pane" });
-
-  const queue = pane.createDiv({ cls: "wd-block" });
-  queue.createDiv({
-    cls: "wd-eyebrow",
-    text: `Queue — ${pipeline.unprocessed.length} unprocessed`,
-  });
-  if (pipeline.unprocessed.length === 0) {
-    queue.createDiv({ cls: "wd-note", text: "Every raw source has a summary page." });
-  }
-  for (const file of pipeline.unprocessed) {
-    const row = queue.createDiv({ cls: "wd-row" });
-    const main = row.createDiv({ cls: "wd-row-main" });
-    main.createDiv({ cls: "wd-row-name", text: file.name });
-    main.createDiv({ cls: "wd-mono", text: formatBytes(file.size) });
-    const btn = row.createEl("button", { cls: "wd-btn" });
-    addIcon(btn, "terminal");
-    btn.createSpan({ text: "Ingest" });
-    if (gate && !gate.ok) {
-      btn.disabled = true;
-      btn.setAttr("title", gate.reason);
-    } else {
-      btn.addEventListener("click", () => onIngest(file, row));
-    }
-  }
-
-  const done = pane.createDiv({ cls: "wd-block" });
-  done.createDiv({ cls: "wd-eyebrow", text: `Ingested — ${pipeline.processed.length}` });
-  const list = done.createDiv({ cls: "wd-done" });
-  for (const file of pipeline.processed) {
-    const row = list.createDiv({ cls: "wd-done-row" });
-    const date = file.page.name.slice(0, 10);
-    row.createSpan({ cls: "wd-mono", text: /^\d{4}-\d{2}-\d{2}$/.test(date) ? date : "—" });
-    row.createSpan({ cls: "wd-done-name", text: file.name.replace(/\.md$/, "") });
-  }
-
-  if (pipeline.orphaned.length > 0) {
-    const orphan = pane.createDiv({ cls: "wd-block" });
-    orphan.createDiv({
-      cls: "wd-eyebrow",
-      text: `Orphaned — ${pipeline.orphaned.length}`,
-    });
-    for (const page of pipeline.orphaned) {
-      const row = orphan.createDiv({ cls: "wd-done-row" });
-      row.createSpan({ cls: "wd-done-name", text: page.name });
-      row.createSpan({ cls: "wd-mono", text: page.rawPath || "no raw path" });
-    }
-  }
-
-  return pane;
-}
-
-const STATUS_ICON = { covered: "check", partial: "minus", absent: "x" };
-
-function renderRightPane(container, parsed, reconciled, cfg) {
-  const pane = container.createDiv({ cls: "wd-pane" });
-  pane.createDiv({ cls: "wd-eyebrow", text: "Coverage — by station" });
-
-  const meta = parsed.meta || {};
-  pane.createDiv({
-    cls: "wd-mono",
-    text: `${meta.script || "no script recorded"} — last synced ${meta.lastSynced || "never"}`,
-  });
-  if (cfg && cfg.script && meta.script && meta.script !== cfg.script) {
-    pane.createDiv({
-      cls: "wd-error",
-      text: `This dashboard is configured for ${cfg.script}, but the coverage file tracks ${meta.script}.`,
-    });
-  }
-
-  const counts = { covered: 0, partial: 0, absent: 0 };
-  for (const row of parsed.rows) counts[row.status] += 1;
-  const total = parsed.rows.length;
-
-  if (total > 0) {
-    const meter = pane.createDiv({ cls: "wd-meter" });
-    meter.setAttr("role", "img");
-    meter.setAttr(
-      "aria-label",
-      `Coverage: ${counts.covered} covered, ${counts.partial} partial, ${counts.absent} absent of ${total}`
-    );
-    for (const key of ["covered", "partial", "absent"]) {
-      if (counts[key] === 0) continue;
-      const seg = meter.createEl("i", { cls: `wd-seg-${key}` });
-      seg.style.flex = String(counts[key]);
-    }
-    const key = pane.createDiv({ cls: "wd-key" });
-    for (const name of ["covered", "partial", "absent"]) {
-      const span = key.createSpan();
-      span.createEl("i", { cls: `wd-seg-${name}` });
-      span.createSpan({ text: `${counts[name]} ${name}` });
-    }
-  }
-
-  if (parsed.errors.length > 0) {
-    const box = pane.createDiv({ cls: "wd-error" });
-    box.createDiv({ text: `${parsed.errors.length} unparseable row(s):` });
-    for (const e of parsed.errors) {
-      box.createDiv({ cls: "wd-mono", text: `line ${e.line} — ${e.reason}` });
-    }
-  }
-
-  const table = pane.createEl("table", { cls: "wd-tbl" });
-  const head = table.createEl("thead").createEl("tr");
-  for (const h of ["Concept", "Status", "Station", ""]) head.createEl("th", { text: h });
-  const body = table.createEl("tbody");
-
-  for (const group of groupByStation(parsed.rows)) {
-    const gr = body.createEl("tr", { cls: "wd-grp" });
-    gr.createEl("td", { attr: { colspan: "4" }, text: `${group.station} — ${group.rows.length}` });
-    for (const row of group.rows) {
-      const tr = body.createEl("tr");
-      // Obsidian's click handler resolves internal links via data-href, so both
-      // attributes are required for the link to open the concept page.
-      tr.createEl("td").createEl("a", {
-        cls: "internal-link",
-        text: row.concept,
-        attr: { href: row.concept, "data-href": row.concept },
-      });
-      const stat = tr.createEl("td").createSpan({ cls: `wd-stat wd-${row.status}` });
-      addIcon(stat, STATUS_ICON[row.status]);
-      stat.createSpan({ text: row.status });
-      tr.createEl("td", { cls: "wd-mono", text: row.stations.join(", ") || "—" });
-      tr.createEl("td", { cls: "wd-pin", text: row.pinned ? "pinned" : "" });
-    }
-  }
-
-  if (reconciled.unsynced.length > 0) {
-    const box = pane.createDiv({ cls: "wd-block" });
-    box.createDiv({ cls: "wd-eyebrow", text: `Unsynced — ${reconciled.unsynced.length}` });
-    box.createDiv({
-      cls: "wd-note",
-      text: `Concept pages with no row. Run "sync script coverage": ${reconciled.unsynced.join(", ")}`,
-    });
-  }
-
-  if (reconciled.stale.length > 0) {
-    const box = pane.createDiv({ cls: "wd-block" });
-    box.createDiv({ cls: "wd-eyebrow", text: `Stale — ${reconciled.stale.length}` });
-    box.createDiv({
-      cls: "wd-note",
-      text: `Rows whose concept page is gone: ${reconciled.stale.map((r) => r.concept).join(", ")}`,
-    });
-  }
-
-  return pane;
-}
-
-class WebinarDashPlugin extends PluginBase {
-  async onload() {
-    this.running = new Set();
-    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
-      const cfg = parseConfig(source);
-      const root = el.createDiv({ cls: "webinar-dash" });
-      await this.renderAll(root, cfg);
-    });
-  }
-
-  async renderAll(root, cfg) {
-    root.empty();
-    const gate = this.spawnGate();
-    const refresh = () => { this.renderAll(root, cfg); };
-
-    const bar = root.createDiv({ cls: "wd-bar" });
-    const sync = bar.createEl("button", { cls: "wd-btn wd-btn-ghost" });
-    addIcon(sync, "refresh");
-    sync.createSpan({ text: "Sync coverage" });
-    if (!gate.ok) {
-      sync.disabled = true;
-      sync.setAttr("title", gate.reason);
-    } else {
-      sync.addEventListener("click", () =>
-        this.runClaude("__sync__", "sync script coverage", bar, "sync coverage", refresh)
-      );
-    }
-
-    const grid = root.createDiv({ cls: "wd-grid" });
-
-    try {
-      const pipeline = await this.readPipeline(cfg);
-      renderLeftPane(grid, pipeline, (file, rowEl) => this.runIngest(file, rowEl, refresh), gate);
-    } catch (err) {
-      grid.createDiv({ cls: "wd-error", text: `Pipeline failed: ${err.message}` });
-    }
-
-    try {
-      const { parsed, reconciled } = await this.readCoverage(cfg);
-      renderRightPane(grid, parsed, reconciled, cfg);
-    } catch (err) {
-      grid.createDiv({ cls: "wd-error", text: `Coverage failed: ${err.message}` });
-    }
-  }
-
-  async readCoverage(cfg) {
-    const file = this.app.vault.getAbstractFileByPath(cfg.coverage);
-    if (!file) throw new Error(`no coverage file at ${cfg.coverage}`);
-    const parsed = parseCoverageTable(await this.app.vault.cachedRead(file));
-
-    const conceptDir = cfg.conceptDir.replace(/\/+$/, "") + "/";
-    const conceptNames = this.app.vault
-      .getFiles()
-      .filter((f) => f.path.startsWith(conceptDir) && f.extension === "md")
-      .map((f) => f.basename);
-
-    return { parsed, reconciled: reconcileConcepts(parsed.rows, conceptNames) };
-  }
-
-  async readPipeline(cfg) {
-    const all = this.app.vault.getFiles();
-    const rawDir = cfg.rawDir.replace(/\/+$/, "") + "/";
-    const wikiDir = cfg.wikiSourceDir.replace(/\/+$/, "") + "/";
-
-    // CLAUDE.md documents raw/sources as holding "markdown/text/pdf exports".
-    // Allow-listing only "md" would hide the others from the queue with no
-    // warning — the same silent-invisibility failure this dashboard exists to
-    // remove. Wiki source pages below stay markdown-only; those really are .md.
-    const rawFiles = all
-      .filter((f) => f.path.startsWith(rawDir) && RAW_EXTENSIONS.has(f.extension))
-      .map((f) => ({ path: f.path, name: f.name, size: f.stat.size }));
-
-    const pageFiles = all.filter((f) => f.path.startsWith(wikiDir) && f.extension === "md");
-    const sourcePages = [];
-    for (const f of pageFiles) {
-      const text = await this.app.vault.cachedRead(f);
-      sourcePages.push({ path: f.path, name: f.name, rawPath: extractRawPath(text) });
-    }
-
-    return derivePipeline({ rawFiles, sourcePages });
-  }
-
-  vaultPath() {
-    const adapter = this.app.vault.adapter;
-    if (typeof adapter.getBasePath === "function") return adapter.getBasePath();
-    return null;
-  }
-
-  spawnGate() {
-    if (!this.vaultPath()) {
-      return { ok: false, reason: "Needs desktop Obsidian." };
-    }
-    try {
-      require("child_process");
-    } catch (_) {
-      return { ok: false, reason: "child_process unavailable — needs desktop Obsidian." };
-    }
-    return { ok: true, reason: "" };
-  }
-
-  // One subprocess runner for every button. `key` is what makes a run unique in
-  // `this.running` — a file path for ingest, a constant for sync — so the guard
-  // survives the row re-renders that a DOM-scoped guard could not.
-  runClaude(key, prompt, hostEl, label, onSuccess) {
-    const Notice = OB ? OB.Notice : null;
-    const notify = (msg) => { if (Notice) new Notice(msg); };
-
-    if (this.running.has(key)) {
-      notify(`Already running: ${label}.`);
-      return;
-    }
-
-    const gate = this.spawnGate();
-    if (!gate.ok) {
-      notify(gate.reason);
-      return;
-    }
-    const { spawn } = require("child_process");
-    const base = this.vaultPath();
-
-    // A retry reuses the same host element. Clear the previous run's status and
-    // output so they are replaced rather than stacked.
-    hostEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
-
-    const button = hostEl.querySelector("button");
-    if (button) button.disabled = true;
-    this.running.add(key);
-
-    const status = hostEl.createSpan({ cls: "wd-status wd-status-running", text: "running 0s" });
-    const output = hostEl.createDiv({ cls: "wd-output", text: "" });
-    const started = Date.now();
-    const timer = window.setInterval(() => {
-      status.setText(`running ${Math.round((Date.now() - started) / 1000)}s`);
-    }, 1000);
-    this.registerInterval(timer);
-
-    // Bounded as it accumulates, not only when displayed.
-    let buffered = "";
-    const append = (text) => {
-      buffered = (buffered + text).slice(-8000);
-      output.setText(buffered);
-      output.scrollTop = output.scrollHeight;
-    };
-
-    // Node can emit both `error` and `close` for one failure. `settled` keeps the
-    // first, more specific message instead of letting `exit null` overwrite it.
-    let settled = false;
-    const finish = (cls, text) => {
-      if (settled) return;
-      settled = true;
-      window.clearInterval(timer);
-      this.running.delete(key);
-      status.className = `wd-status ${cls}`;
-      status.setText(text);
-      if (button) button.disabled = false;
-    };
-
-    let child;
-    try {
-      // See the C1 comment in runIngest: no shell, deliberately.
-      child = spawn("claude", ["-p", prompt], { cwd: base });
-    } catch (err) {
-      append(`\nCould not start claude: ${err.message}`);
-      finish("wd-status-failed", "failed");
-      return;
-    }
-
-    child.stdout.setEncoding("utf8");
-    child.stderr.setEncoding("utf8");
-    child.stdout.on("data", append);
-    child.stderr.on("data", append);
-
-    child.on("error", (err) => {
-      append(`\nCould not start claude: ${err.message}\nIs it on PATH?`);
-      finish("wd-status-failed", "failed");
-    });
-
-    child.on("close", (code) => {
-      const secs = Math.round((Date.now() - started) / 1000);
-      if (code === 0) {
-        finish("wd-status-done", `done in ${secs}s`);
-        if (onSuccess) onSuccess();
-      } else {
-        finish("wd-status-failed", `failed - exit ${code}`);
-      }
-    });
-  }
-
-  runIngest(file, rowEl, onSuccess) {
-    const Notice = OB ? OB.Notice : null;
-
-    if (!isSafeFilename(file.name)) {
-      rowEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
-      rowEl.createDiv({
-        cls: "wd-output",
-        text: `Refused: "${file.name}" contains a character that is unsafe to pass to a shell.`,
-      });
-      return;
-    }
-
-    // `claude` resolves to a real .exe here, so libuv finds it via PATH and
-    // PATHEXT with no shell involved. Do NOT add `shell: true`: with a shell,
-    // Node concatenates argv without quoting and cmd.exe re-tokenizes, so the
-    // prompt arrives as two arguments and the filename is silently discarded.
-    // Verified: shell:true yields ["-p","ingest","Webinar script.md"], where the
-    // program sees only "ingest". If a .cmd shim ever needs supporting, use the
-    // single-string form spawn(`claude -p "ingest ${name}"`, { shell: true }).
-    this.runClaude(
-      file.path,
-      `ingest "${file.name}"`,
-      rowEl,
-      file.name,
-      () => {
-        if (Notice) new Notice(`Ingested ${file.name}.`);
-        if (onSuccess) onSuccess();
-      }
-    );
-  }
-}
-
-module.exports = WebinarDashPlugin;
-module.exports.default = WebinarDashPlugin;
-module.exports.__test__ = {
-  parseConfig, extractRawPath, derivePipeline,
-  parseCoverageTable, groupByStation, reconcileConcepts,
-  STATIONS, DEFAULTS, isSafeFilename,
-};
+"use strict";
+
+// Obsidian injects its own module resolver. Under plain `node --test` it is
+// absent, so guard the require and fall back to an empty base class. This is
+// what keeps the pure helpers below unit-testable without an Obsidian runtime.
+let OB = null;
+try {
+  OB = require("obsidian");
+} catch (_) {
+  OB = null;
+}
+const PluginBase = OB ? OB.Plugin : class {};
+
+const STATIONS = ["Chat box", "ReAct", "Tools", "Memory", "Skills", "Process", "OS"];
+
+const DEFAULTS = {
+  script: "raw/sources/Webinar script.md",
+  coverage: "wiki/script-coverage.md",
+  rawDir: "raw/sources",
+  wikiSourceDir: "wiki/sources",
+  conceptDir: "wiki/concepts",
+};
+
+const RAW_PATH_RE = /^\s*-\s*\*\*Raw path:\*\*\s*`([^`]+)`/m;
+
+function extractRawPath(text) {
+  const m = String(text).match(RAW_PATH_RE);
+  return m ? m[1].trim() : null;
+}
+
+function derivePipeline({ rawFiles, sourcePages }) {
+  // First claim on a raw path wins. A later page claiming the same file is a
+  // duplicate claim — real catalog drift — and joins `orphaned` rather than
+  // being silently dropped. `orphaned` therefore means "source page not paired
+  // with a raw file", whatever the reason.
+  const claimed = new Map();
+  const duplicates = [];
+  for (const page of sourcePages) {
+    if (!page.rawPath) continue;
+    if (claimed.has(page.rawPath)) duplicates.push(page);
+    else claimed.set(page.rawPath, page);
+  }
+
+  const processed = [];
+  const unprocessed = [];
+  for (const file of rawFiles) {
+    const page = claimed.get(file.path);
+    if (page) processed.push(Object.assign({}, file, { page }));
+    else unprocessed.push(file);
+  }
+
+  // One pass over sourcePages, so a page appears in `orphaned` at most once no
+  // matter how many of the three reasons apply to it. Concatenating a separate
+  // duplicates array here would double-count a losing claimant whose shared raw
+  // path is also missing from disk.
+  const rawPaths = new Set(rawFiles.map((f) => f.path));
+  const duplicateSet = new Set(duplicates);
+  const orphaned = sourcePages.filter(
+    (p) => duplicateSet.has(p) || !p.rawPath || !rawPaths.has(p.rawPath)
+  );
+
+  unprocessed.sort((a, b) => a.name.localeCompare(b.name));
+  processed.sort((a, b) => b.page.name.localeCompare(a.page.name));
+
+  return { processed, unprocessed, orphaned };
+}
+
+const VALID_STATUS = new Set(["covered", "partial", "absent"]);
+const SEPARATOR_RE = /^\|?\s*:?-{2,}/;
+
+function parseCoverageTable(markdown) {
+  const text = String(markdown);
+  const meta = { script: null, lastSynced: null };
+
+  const scriptM = text.match(/^\s*-\s*\*\*Script:\*\*\s*`([^`]+)`/m);
+  if (scriptM) meta.script = scriptM[1].trim();
+  const syncM = text.match(/^\s*-\s*\*\*Last synced:\*\*\s*(\S+)/m);
+  if (syncM) meta.lastSynced = syncM[1].trim();
+
+  const rows = [];
+  const errors = [];
+  let inTable = false;
+
+  text.split(/\r?\n/).forEach((line, i) => {
+    const t = line.trim();
+    if (!t.startsWith("|")) {
+      inTable = false;
+      return;
+    }
+    if (SEPARATOR_RE.test(t)) {
+      inTable = true;
+      return;
+    }
+    if (!inTable) return;
+
+    const body = t.endsWith("|") ? t.slice(1, -1) : t.slice(1);
+    // Split on unescaped pipes only. Obsidian escapes the pipe of a piped
+    // wikilink inside a table cell as `\|`, and a plain split("|") tears
+    // `[[harness\|alias]]` into two cells, shifting every later column left.
+    const cells = body.split(/(?<!\\)\|/).map((c) => c.trim());
+
+    if (cells.length < 3) {
+      errors.push({ line: i + 1, text: t, reason: "expected at least 3 columns" });
+      return;
+    }
+
+    const status = cells[1].toLowerCase();
+    if (!VALID_STATUS.has(status)) {
+      errors.push({ line: i + 1, text: t, reason: `invalid status "${cells[1]}"` });
+      return;
+    }
+
+    // Capture up to the first ] | or backslash. Obsidian escapes the pipe in a
+    // piped wikilink inside a table cell, so the raw cell reads [[name\|alias]] —
+    // excluding the backslash is what keeps the trailing "\" out of the name.
+    const linkM = cells[0].match(/\[\[([^\]|\\]+)/);
+    const stationCell = cells[2];
+
+    rows.push({
+      concept: linkM ? linkM[1].trim() : cells[0],
+      status,
+      stations:
+        stationCell === "—" || stationCell === "-" || stationCell === ""
+          ? []
+          : stationCell.split(",").map((s) => s.trim()).filter(Boolean),
+      pinned: (cells[3] || "").toLowerCase() === "yes",
+      line: i + 1,
+    });
+  });
+
+  return { meta, rows, errors };
+}
+
+function groupByStation(rows) {
+  const order = ["All stations", ...STATIONS, "No station"];
+  const buckets = new Map(order.map((k) => [k, []]));
+
+  for (const row of rows) {
+    let key;
+    if (row.stations.includes("all")) key = "All stations";
+    else if (row.stations.length === 0) key = "No station";
+    else key = row.stations[0];
+    if (!buckets.has(key)) buckets.set(key, []);
+    buckets.get(key).push(row);
+  }
+
+  const known = order.filter((k) => buckets.get(k).length > 0);
+  const unknown = [...buckets.keys()].filter((k) => !order.includes(k) && buckets.get(k).length > 0);
+  return [...known, ...unknown].map((station) => ({ station, rows: buckets.get(station) }));
+}
+
+function reconcileConcepts(rows, conceptNames) {
+  const named = new Set(conceptNames);
+  const rowed = new Set(rows.map((r) => r.concept));
+  return {
+    rows,
+    unsynced: conceptNames.filter((n) => !rowed.has(n)).sort(),
+    stale: rows.filter((r) => !named.has(r.concept)),
+  };
+}
+
+// Defence in depth. The spawn path no longer uses a shell, so nothing here is
+// load-bearing against injection today — but the guard costs nothing and would
+// still hold if the single-string `shell: true` form is ever needed for a .cmd
+// shim. `^` is cmd.exe's escape character; control characters are rejected
+// because a NUL byte makes spawn() throw synchronously.
+//
+// Apostrophes, spaces, cyrillic, em dashes and `!` stay allowed — they appear in
+// real filenames in this vault.
+const UNSAFE_CHARS = /["`$&|;<>%^\u0000-\u001f\u007f]/;
+
+function isSafeFilename(name) {
+  if (typeof name !== "string" || name.length === 0) return false;
+  if (UNSAFE_CHARS.test(name)) return false;
+  if (name.includes("..")) return false;
+  return true;
+}
+
+function parseConfig(source) {
+  const cfg = Object.assign({}, DEFAULTS);
+  for (const line of String(source).split(/\r?\n/)) {
+    const m = line.match(/^\s*([A-Za-z][A-Za-z0-9_]*)\s*:\s*(.+?)\s*$/);
+    if (m && Object.prototype.hasOwnProperty.call(DEFAULTS, m[1])) {
+      cfg[m[1]] = m[2];
+    }
+  }
+  return cfg;
+}
+
+// Source types CLAUDE.md documents for raw/sources. Anything else in that
+// folder is not a source and stays out of the queue.
+const RAW_EXTENSIONS = new Set(["md", "txt", "pdf"]);
+
+const ICONS = {
+  terminal: "M4 17l6-6-6-6M12 19h8",
+  check: "M20 6 9 17l-5-5",
+  minus: "M5 12h14",
+  x: "M18 6 6 18M6 6l12 12",
+  refresh: "M3 12a9 9 0 0 1 9-9 9 9 0 0 1 6.7 3L21 8M21 3v5h-5 M21 12a9 9 0 0 1-9 9 9 9 0 0 1-6.7-3L3 16M3 21v-5h5",
+};
+
+// Built with createElementNS rather than Obsidian's createSvg helper, whose
+// availability varies by version. This works on any Obsidian build.
+const SVG_NS = "http://www.w3.org/2000/svg";
+
+function addIcon(parent, name) {
+  const svg = document.createElementNS(SVG_NS, "svg");
+  svg.setAttribute("viewBox", "0 0 24 24");
+  svg.setAttribute("fill", "none");
+  svg.setAttribute("aria-hidden", "true");
+  const path = document.createElementNS(SVG_NS, "path");
+  path.setAttribute("d", ICONS[name]);
+  path.setAttribute("stroke", "currentColor");
+  path.setAttribute("stroke-width", "1.75");
+  path.setAttribute("stroke-linecap", "round");
+  path.setAttribute("stroke-linejoin", "round");
+  svg.appendChild(path);
+  parent.appendChild(svg);
+  return svg;
+}
+
+function formatBytes(n) {
+  return `${n.toLocaleString("en-US")} B`;
+}
+
+function renderLeftPane(container, pipeline, onIngest, gate) {
+  const pane = container.createDiv({ cls: "wd-pane" });
+
+  const queue = pane.createDiv({ cls: "wd-block" });
+  queue.createDiv({
+    cls: "wd-eyebrow",
+    text: `Queue — ${pipeline.unprocessed.length} unprocessed`,
+  });
+  if (pipeline.unprocessed.length === 0) {
+    queue.createDiv({ cls: "wd-note", text: "Every raw source has a summary page." });
+  }
+  for (const file of pipeline.unprocessed) {
+    const row = queue.createDiv({ cls: "wd-row" });
+    const main = row.createDiv({ cls: "wd-row-main" });
+    main.createDiv({ cls: "wd-row-name", text: file.name });
+    main.createDiv({ cls: "wd-mono", text: formatBytes(file.size) });
+    const btn = row.createEl("button", { cls: "wd-btn" });
+    addIcon(btn, "terminal");
+    btn.createSpan({ text: "Ingest" });
+    if (gate && !gate.ok) {
+      btn.disabled = true;
+      btn.setAttr("title", gate.reason);
+    } else {
+      btn.addEventListener("click", () => onIngest(file, row));
+    }
+  }
+
+  const done = pane.createDiv({ cls: "wd-block" });
+  done.createDiv({ cls: "wd-eyebrow", text: `Ingested — ${pipeline.processed.length}` });
+  const list = done.createDiv({ cls: "wd-done" });
+  for (const file of pipeline.processed) {
+    const row = list.createDiv({ cls: "wd-done-row" });
+    const date = file.page.name.slice(0, 10);
+    row.createSpan({ cls: "wd-mono", text: /^\d{4}-\d{2}-\d{2}$/.test(date) ? date : "—" });
+    row.createSpan({ cls: "wd-done-name", text: file.name.replace(/\.md$/, "") });
+  }
+
+  if (pipeline.orphaned.length > 0) {
+    const orphan = pane.createDiv({ cls: "wd-block" });
+    orphan.createDiv({
+      cls: "wd-eyebrow",
+      text: `Orphaned — ${pipeline.orphaned.length}`,
+    });
+    for (const page of pipeline.orphaned) {
+      const row = orphan.createDiv({ cls: "wd-done-row" });
+      row.createSpan({ cls: "wd-done-name", text: page.name });
+      row.createSpan({ cls: "wd-mono", text: page.rawPath || "no raw path" });
+    }
+  }
+
+  return pane;
+}
+
+const STATUS_ICON = { covered: "check", partial: "minus", absent: "x" };
+
+function renderRightPane(container, parsed, reconciled, cfg) {
+  const pane = container.createDiv({ cls: "wd-pane" });
+  pane.createDiv({ cls: "wd-eyebrow", text: "Coverage — by station" });
+
+  const meta = parsed.meta || {};
+  pane.createDiv({
+    cls: "wd-mono",
+    text: `${meta.script || "no script recorded"} — last synced ${meta.lastSynced || "never"}`,
+  });
+  if (cfg && cfg.script && meta.script && meta.script !== cfg.script) {
+    pane.createDiv({
+      cls: "wd-error",
+      text: `This dashboard is configured for ${cfg.script}, but the coverage file tracks ${meta.script}.`,
+    });
+  }
+
+  const counts = { covered: 0, partial: 0, absent: 0 };
+  for (const row of parsed.rows) counts[row.status] += 1;
+  const total = parsed.rows.length;
+
+  if (total > 0) {
+    const meter = pane.createDiv({ cls: "wd-meter" });
+    meter.setAttr("role", "img");
+    meter.setAttr(
+      "aria-label",
+      `Coverage: ${counts.covered} covered, ${counts.partial} partial, ${counts.absent} absent of ${total}`
+    );
+    for (const key of ["covered", "partial", "absent"]) {
+      if (counts[key] === 0) continue;
+      const seg = meter.createEl("i", { cls: `wd-seg-${key}` });
+      seg.style.flex = String(counts[key]);
+    }
+    const key = pane.createDiv({ cls: "wd-key" });
+    for (const name of ["covered", "partial", "absent"]) {
+      const span = key.createSpan();
+      span.createEl("i", { cls: `wd-seg-${name}` });
+      span.createSpan({ text: `${counts[name]} ${name}` });
+    }
+  }
+
+  if (parsed.errors.length > 0) {
+    const box = pane.createDiv({ cls: "wd-error" });
+    box.createDiv({ text: `${parsed.errors.length} unparseable row(s):` });
+    for (const e of parsed.errors) {
+      box.createDiv({ cls: "wd-mono", text: `line ${e.line} — ${e.reason}` });
+    }
+  }
+
+  const table = pane.createEl("table", { cls: "wd-tbl" });
+  const head = table.createEl("thead").createEl("tr");
+  for (const h of ["Concept", "Status", "Station", ""]) head.createEl("th", { text: h });
+  const body = table.createEl("tbody");
+
+  for (const group of groupByStation(parsed.rows)) {
+    const gr = body.createEl("tr", { cls: "wd-grp" });
+    gr.createEl("td", { attr: { colspan: "4" }, text: `${group.station} — ${group.rows.length}` });
+    for (const row of group.rows) {
+      const tr = body.createEl("tr");
+      // Obsidian's click handler resolves internal links via data-href, so both
+      // attributes are required for the link to open the concept page.
+      tr.createEl("td").createEl("a", {
+        cls: "internal-link",
+        text: row.concept,
+        attr: { href: row.concept, "data-href": row.concept },
+      });
+      const stat = tr.createEl("td").createSpan({ cls: `wd-stat wd-${row.status}` });
+      addIcon(stat, STATUS_ICON[row.status]);
+      stat.createSpan({ text: row.status });
+      tr.createEl("td", { cls: "wd-mono", text: row.stations.join(", ") || "—" });
+      tr.createEl("td", { cls: "wd-pin", text: row.pinned ? "pinned" : "" });
+    }
+  }
+
+  if (reconciled.unsynced.length > 0) {
+    const box = pane.createDiv({ cls: "wd-block" });
+    box.createDiv({ cls: "wd-eyebrow", text: `Unsynced — ${reconciled.unsynced.length}` });
+    box.createDiv({
+      cls: "wd-note",
+      text: `Concept pages with no row. Run "sync script coverage": ${reconciled.unsynced.join(", ")}`,
+    });
+  }
+
+  if (reconciled.stale.length > 0) {
+    const box = pane.createDiv({ cls: "wd-block" });
+    box.createDiv({ cls: "wd-eyebrow", text: `Stale — ${reconciled.stale.length}` });
+    box.createDiv({
+      cls: "wd-note",
+      text: `Rows whose concept page is gone: ${reconciled.stale.map((r) => r.concept).join(", ")}`,
+    });
+  }
+
+  return pane;
+}
+
+class WebinarDashPlugin extends PluginBase {
+  async onload() {
+    this.running = new Set();
+    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
+      const cfg = parseConfig(source);
+      const root = el.createDiv({ cls: "webinar-dash" });
+      await this.renderAll(root, cfg);
+    });
+  }
+
+  async renderAll(root, cfg) {
+    root.empty();
+    const gate = this.spawnGate();
+    const refresh = () => { this.renderAll(root, cfg); };
+
+    const bar = root.createDiv({ cls: "wd-bar" });
+    const sync = bar.createEl("button", { cls: "wd-btn wd-btn-ghost" });
+    addIcon(sync, "refresh");
+    sync.createSpan({ text: "Sync coverage" });
+    if (!gate.ok) {
+      sync.disabled = true;
+      sync.setAttr("title", gate.reason);
+    } else {
+      sync.addEventListener("click", () =>
+        this.runClaude("__sync__", "sync script coverage", bar, "sync coverage", refresh)
+      );
+    }
+
+    const grid = root.createDiv({ cls: "wd-grid" });
+
+    try {
+      const pipeline = await this.readPipeline(cfg);
+      renderLeftPane(grid, pipeline, (file, rowEl) => this.runIngest(file, rowEl, refresh), gate);
+    } catch (err) {
+      grid.createDiv({ cls: "wd-error", text: `Pipeline failed: ${err.message}` });
+    }
+
+    try {
+      const { parsed, reconciled } = await this.readCoverage(cfg);
+      renderRightPane(grid, parsed, reconciled, cfg);
+    } catch (err) {
+      grid.createDiv({ cls: "wd-error", text: `Coverage failed: ${err.message}` });
+    }
+  }
+
+  async readCoverage(cfg) {
+    const file = this.app.vault.getAbstractFileByPath(cfg.coverage);
+    if (!file) throw new Error(`no coverage file at ${cfg.coverage}`);
+    const parsed = parseCoverageTable(await this.app.vault.cachedRead(file));
+
+    const conceptDir = cfg.conceptDir.replace(/\/+$/, "") + "/";
+    const conceptNames = this.app.vault
+      .getFiles()
+      .filter((f) => f.path.startsWith(conceptDir) && f.extension === "md")
+      .map((f) => f.basename);
+
+    return { parsed, reconciled: reconcileConcepts(parsed.rows, conceptNames) };
+  }
+
+  async readPipeline(cfg) {
+    const all = this.app.vault.getFiles();
+    const rawDir = cfg.rawDir.replace(/\/+$/, "") + "/";
+    const wikiDir = cfg.wikiSourceDir.replace(/\/+$/, "") + "/";
+
+    // CLAUDE.md documents raw/sources as holding "markdown/text/pdf exports".
+    // Allow-listing only "md" would hide the others from the queue with no
+    // warning — the same silent-invisibility failure this dashboard exists to
+    // remove. Wiki source pages below stay markdown-only; those really are .md.
+    const rawFiles = all
+      .filter((f) => f.path.startsWith(rawDir) && RAW_EXTENSIONS.has(f.extension))
+      .map((f) => ({ path: f.path, name: f.name, size: f.stat.size }));
+
+    const pageFiles = all.filter((f) => f.path.startsWith(wikiDir) && f.extension === "md");
+    const sourcePages = [];
+    for (const f of pageFiles) {
+      const text = await this.app.vault.cachedRead(f);
+      sourcePages.push({ path: f.path, name: f.name, rawPath: extractRawPath(text) });
+    }
+
+    return derivePipeline({ rawFiles, sourcePages });
+  }
+
+  vaultPath() {
+    const adapter = this.app.vault.adapter;
+    if (typeof adapter.getBasePath === "function") return adapter.getBasePath();
+    return null;
+  }
+
+  spawnGate() {
+    if (!this.vaultPath()) {
+      return { ok: false, reason: "Needs desktop Obsidian." };
+    }
+    try {
+      require("child_process");
+    } catch (_) {
+      return { ok: false, reason: "child_process unavailable — needs desktop Obsidian." };
+    }
+    return { ok: true, reason: "" };
+  }
+
+  // One subprocess runner for every button. `key` is what makes a run unique in
+  // `this.running` — a file path for ingest, a constant for sync — so the guard
+  // survives the row re-renders that a DOM-scoped guard could not.
+  runClaude(key, prompt, hostEl, label, onSuccess) {
+    const Notice = OB ? OB.Notice : null;
+    const notify = (msg) => { if (Notice) new Notice(msg); };
+
+    if (this.running.has(key)) {
+      notify(`Already running: ${label}.`);
+      return;
+    }
+
+    const gate = this.spawnGate();
+    if (!gate.ok) {
+      notify(gate.reason);
+      return;
+    }
+    const { spawn } = require("child_process");
+    const base = this.vaultPath();
+
+    // A retry reuses the same host element. Clear the previous run's status and
+    // output so they are replaced rather than stacked.
+    hostEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
+
+    const button = hostEl.querySelector("button");
+    if (button) button.disabled = true;
+    this.running.add(key);
+
+    const status = hostEl.createSpan({ cls: "wd-status wd-status-running", text: "running 0s" });
+    const output = hostEl.createDiv({ cls: "wd-output", text: "" });
+    const started = Date.now();
+    const timer = window.setInterval(() => {
+      status.setText(`running ${Math.round((Date.now() - started) / 1000)}s`);
+    }, 1000);
+    this.registerInterval(timer);
+
+    // Bounded as it accumulates, not only when displayed.
+    let buffered = "";
+    const append = (text) => {
+      buffered = (buffered + text).slice(-8000);
+      output.setText(buffered);
+      output.scrollTop = output.scrollHeight;
+    };
+
+    // Node can emit both `error` and `close` for one failure. `settled` keeps the
+    // first, more specific message instead of letting `exit null` overwrite it.
+    let settled = false;
+    const finish = (cls, text) => {
+      if (settled) return;
+      settled = true;
+      window.clearInterval(timer);
+      this.running.delete(key);
+      status.className = `wd-status ${cls}`;
+      status.setText(text);
+      if (button) button.disabled = false;
+    };
+
+    let child;
+    try {
+      // See the C1 comment in runIngest: no shell, deliberately.
+      child = spawn("claude", ["-p", prompt], { cwd: base });
+    } catch (err) {
+      append(`\nCould not start claude: ${err.message}`);
+      finish("wd-status-failed", "failed");
+      return;
+    }
+
+    child.stdout.setEncoding("utf8");
+    child.stderr.setEncoding("utf8");
+    child.stdout.on("data", append);
+    child.stderr.on("data", append);
+
+    child.on("error", (err) => {
+      append(`\nCould not start claude: ${err.message}\nIs it on PATH?`);
+      finish("wd-status-failed", "failed");
+    });
+
+    child.on("close", (code) => {
+      const secs = Math.round((Date.now() - started) / 1000);
+      if (code === 0) {
+        finish("wd-status-done", `done in ${secs}s`);
+        if (onSuccess) onSuccess();
+      } else {
+        finish("wd-status-failed", `failed - exit ${code}`);
+      }
+    });
+  }
+
+  runIngest(file, rowEl, onSuccess) {
+    const Notice = OB ? OB.Notice : null;
+
+    if (!isSafeFilename(file.name)) {
+      rowEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
+      rowEl.createDiv({
+        cls: "wd-output",
+        text: `Refused: "${file.name}" contains a character that is unsafe to pass to a shell.`,
+      });
+      return;
+    }
+
+    // `claude` resolves to a real .exe here, so libuv finds it via PATH and
+    // PATHEXT with no shell involved. Do NOT add `shell: true`: with a shell,
+    // Node concatenates argv without quoting and cmd.exe re-tokenizes, so the
+    // prompt arrives as two arguments and the filename is silently discarded.
+    // Verified: shell:true yields ["-p","ingest","Webinar script.md"], where the
+    // program sees only "ingest". If a .cmd shim ever needs supporting, use the
+    // single-string form spawn(`claude -p "ingest ${name}"`, { shell: true }).
+    this.runClaude(
+      file.path,
+      `ingest "${file.name}"`,
+      rowEl,
+      file.name,
+      () => {
+        if (Notice) new Notice(`Ingested ${file.name}.`);
+        if (onSuccess) onSuccess();
+      }
+    );
+  }
+}
+
+module.exports = WebinarDashPlugin;
+module.exports.default = WebinarDashPlugin;
+module.exports.__test__ = {
+  parseConfig, extractRawPath, derivePipeline,
+  parseCoverageTable, groupByStation, reconcileConcepts,
+  STATIONS, DEFAULTS, isSafeFilename,
+};
diff --git a/.obsidian/plugins/webinar-dash/manifest.json b/.obsidian/plugins/webinar-dash/manifest.json
index de1ae92..b3322a9 100644
--- a/.obsidian/plugins/webinar-dash/manifest.json
+++ b/.obsidian/plugins/webinar-dash/manifest.json
@@ -1,9 +1,9 @@
-{
-  "id": "webinar-dash",
-  "name": "Webinar dashboard",
-  "version": "0.1.0",
-  "minAppVersion": "1.5.0",
-  "description": "Source pipeline and script coverage for the webinar vault.",
-  "author": "meels",
-  "isDesktopOnly": true
-}
+{
+  "id": "webinar-dash",
+  "name": "Webinar dashboard",
+  "version": "0.1.0",
+  "minAppVersion": "1.5.0",
+  "description": "Source pipeline and script coverage for the webinar vault.",
+  "author": "meels",
+  "isDesktopOnly": true
+}
diff --git a/.obsidian/plugins/webinar-dash/styles.css b/.obsidian/plugins/webinar-dash/styles.css
index 4df15b2..a68c3d6 100644
--- a/.obsidian/plugins/webinar-dash/styles.css
+++ b/.obsidian/plugins/webinar-dash/styles.css
@@ -1,239 +1,239 @@
-.webinar-dash {
-  --wd-ink-000: #ffffff; --wd-ink-050: #f7f7f7; --wd-ink-100: #ececec;
-  --wd-ink-200: #d9d9d9; --wd-ink-400: #8a8a8a; --wd-ink-500: #5e5e5e;
-  --wd-ink-700: #262626; --wd-ink-900: #0a0a0a; --wd-ink-999: #000000;
-  --wd-red-500: #e1261c; --wd-red-600: #c31c14;
-
-  --wd-bg: var(--wd-ink-000);
-  --wd-bg-subtle: var(--wd-ink-050);
-  --wd-fg: var(--wd-ink-999);
-  --wd-fg-2: var(--wd-ink-700);
-  --wd-fg-3: var(--wd-ink-500);
-  --wd-fg-4: var(--wd-ink-400);
-  --wd-border: var(--wd-ink-200);
-  --wd-border-strong: var(--wd-ink-999);
-  --wd-border-subtle: var(--wd-ink-100);
-  --wd-accent: var(--wd-red-500);
-  --wd-accent-press: var(--wd-red-600);
-  --wd-accent-on: #ffffff;
-  --wd-ok: #0a8a3f;
-  --wd-warn: #c68a00;
-  --wd-danger: var(--wd-red-500);
-
-  --wd-mono: "JetBrains Mono", ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
-  --wd-sans: "Inter", ui-sans-serif, system-ui, -apple-system, "Helvetica Neue", Arial, sans-serif;
-
-  font-family: var(--wd-sans);
-  color: var(--wd-fg);
-  display: flex;
-  flex-direction: column;
-  gap: 16px;
-  /* Past roughly this width the two panes stop reading as a pair and table
-     rows get hard to track across. Cap the whole dashboard, not just the
-     grid, so the toolbar stays flush with the panes below it. */
-  max-width: 1600px;
-  /* Makes this element the reference for the @container query further down,
-     so the grid collapses on the pane's width rather than the window's. */
-  container-type: inline-size;
-}
-
-.theme-dark .webinar-dash {
-  --wd-bg: var(--wd-ink-900);
-  --wd-bg-subtle: #161616;
-  --wd-fg: var(--wd-ink-000);
-  --wd-fg-2: var(--wd-ink-200);
-  --wd-fg-3: var(--wd-ink-400);
-  --wd-fg-4: var(--wd-ink-500);
-  --wd-border: var(--wd-ink-700);
-  --wd-border-strong: #b8b8b8;
-  --wd-border-subtle: #161616;
-  /* Accent tracks the installed theme at .obsidian/themes/tesanti/theme.css,
-     whose dark section reads "Black canvas, same signal red": --accent-h/-s/-l
-     are declared once at :root and never redeclared under .theme-dark, and only
-     the hover state lifts (#c31c14 light, #ff4d43 dark). Match that exactly. */
-  --wd-accent: var(--wd-red-500);
-  --wd-accent-press: #ff4d43;
-  --wd-accent-on: #ffffff;
-  /* Status tokens are separate from the brand accent by design-system rule, and
-     here they carry small text in a dense table. #e1261c on near-black is about
-     3.8:1, under AA for small text, so these lift where the accent does not. */
-  --wd-ok: #2fbf6a;
-  --wd-warn: #e0a516;
-  --wd-danger: #ff5c50;
-}
-
-/* ------------------------------------------------------------------
-   Escape Obsidian's readable-line-length cap.
-
-   With "Readable line length" on (the default), Obsidian caps note
-   content at --file-line-width, roughly 700px. That is right for prose
-   and wrong for a two-pane dashboard, which gets squeezed into a column.
-
-   :has() scopes this to the sizer that actually contains a dashboard, so
-   every other note in the vault keeps its readable width. Both selectors
-   are needed: reading view sizes on .markdown-preview-sizer, live preview
-   on .cm-sizer.
-
-   If you would rather not rely on this, the alternatives are Settings ->
-   Editor -> Readable line length (off, but that widens every note), or
-   adding `cssclasses: wide-dash` to the note's frontmatter and swapping
-   the :has() selectors below for `.wide-dash .markdown-preview-sizer`.
-   ------------------------------------------------------------------ */
-.markdown-preview-sizer:has(.webinar-dash),
-.markdown-source-view.mod-cm6 .cm-sizer:has(.webinar-dash) {
-  max-width: none;
-}
-
-.wd-grid {
-  display: grid;
-  grid-template-columns: minmax(0, 38fr) minmax(0, 62fr);
-  gap: 24px;
-}
-.wd-grid > * { min-width: 0; }
-
-.wd-bar {
-  display: flex;
-  align-items: center;
-  gap: 12px;
-  flex-wrap: wrap;
-  border: 1px solid var(--wd-border);
-  border-radius: 4px;
-  padding: 8px 12px;
-  background: var(--wd-bg-subtle);
-}
-
-.wd-btn-ghost {
-  background: transparent;
-  color: var(--wd-fg);
-  border-color: var(--wd-border-strong);
-}
-.wd-btn-ghost:hover {
-  background: var(--wd-bg-muted, var(--wd-bg-subtle));
-  border-color: var(--wd-border-strong);
-}
-
-/* Collapse to one column on the width that actually matters — the pane's,
-   not the window's. A media query measures the window, so a wide window
-   with the dashboard in a narrow split pane would keep two cramped
-   columns. The container query below responds to the pane itself; the
-   media query stays as a floor for a genuinely narrow window. */
-@container (max-width: 820px) {
-  .wd-grid { grid-template-columns: minmax(0, 1fr); }
-}
-
-@media (max-width: 820px) {
-  .wd-grid { grid-template-columns: minmax(0, 1fr); }
-}
-
-.wd-pane { display: flex; flex-direction: column; gap: 20px; }
-
-.wd-eyebrow {
-  font-family: var(--wd-mono); font-size: 11px; line-height: 1;
-  letter-spacing: 0.12em; text-transform: uppercase;
-  color: var(--wd-fg-3); font-weight: 500;
-  display: flex; align-items: center; gap: 8px;
-}
-.wd-eyebrow::before {
-  content: ""; width: 5px; height: 5px; flex: none; background: var(--wd-accent);
-}
-
-.wd-block { display: flex; flex-direction: column; gap: 12px; }
-
-.wd-row {
-  display: flex; align-items: center; gap: 12px; flex-wrap: wrap;
-  border: 1px solid var(--wd-border); border-radius: 4px;
-  padding: 12px 12px 12px 16px; background: var(--wd-bg);
-}
-.wd-row + .wd-row { margin-top: -1px; }
-.wd-row-main { flex: 1; min-width: 0; display: flex; flex-direction: column; gap: 2px; }
-.wd-row-name {
-  font-size: 14px; font-weight: 600;
-  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
-}
-.wd-mono {
-  font-family: var(--wd-mono); font-size: 11px; color: var(--wd-fg-3);
-  font-variant-numeric: tabular-nums;
-}
-
-.wd-btn {
-  display: inline-flex; align-items: center; gap: 6px; flex: none;
-  font-family: var(--wd-sans); font-size: 13px; font-weight: 600;
-  padding: 6px 12px; border-radius: 6px; cursor: pointer;
-  border: 1px solid var(--wd-accent); background: var(--wd-accent);
-  color: var(--wd-accent-on);
-  transition: background 120ms cubic-bezier(0.2, 0, 0, 1);
-}
-.wd-btn:hover { background: var(--wd-accent-press); border-color: var(--wd-accent-press); }
-.wd-btn:focus-visible { outline: 2px solid var(--wd-accent); outline-offset: 2px; }
-.wd-btn[disabled] {
-  opacity: 0.45; cursor: not-allowed;
-  background: transparent; color: var(--wd-fg-3); border-color: var(--wd-border);
-}
-.wd-btn svg { width: 14px; height: 14px; flex: none; }
-
-.wd-done { display: flex; flex-direction: column; }
-.wd-done-row {
-  display: flex; align-items: baseline; gap: 12px; padding: 5px 0;
-  border-bottom: 1px solid var(--wd-border-subtle); font-size: 13px;
-}
-.wd-done-row:last-child { border-bottom: 0; }
-.wd-done-name {
-  color: var(--wd-fg-2);
-  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
-}
-
-.wd-note { font-size: 12px; color: var(--wd-fg-3); }
-.wd-error {
-  font-size: 13px; color: var(--wd-danger);
-  border: 1px solid var(--wd-danger); border-radius: 4px; padding: 12px;
-}
-
-@media (prefers-reduced-motion: reduce) {
-  .webinar-dash * { transition-duration: 0.01ms !important; }
-}
-
-.wd-meter { display: flex; gap: 2px; height: 8px; width: 100%; }
-.wd-meter > i { display: block; height: 100%; }
-.wd-seg-covered { background: var(--wd-ok); }
-.wd-seg-partial { background: var(--wd-warn); }
-.wd-seg-absent { background: var(--wd-danger); }
-
-.wd-key { display: flex; flex-wrap: wrap; gap: 16px; font-size: 12px; color: var(--wd-fg-2); }
-.wd-key > span { display: inline-flex; align-items: center; gap: 6px; }
-.wd-key i { width: 8px; height: 8px; flex: none; }
-
-.wd-tbl { width: 100%; border-collapse: collapse; font-size: 13px; }
-.wd-tbl th {
-  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
-  text-transform: uppercase; color: var(--wd-fg-3); font-weight: 500;
-  text-align: left; padding: 0 12px 8px 0;
-  border-bottom: 1px solid var(--wd-border-strong);
-}
-.wd-tbl td {
-  padding: 6px 12px 6px 0; border-bottom: 1px solid var(--wd-border-subtle);
-  vertical-align: baseline;
-}
-.wd-grp td {
-  padding-top: 16px; border-bottom: 1px solid var(--wd-border-strong);
-  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
-  text-transform: uppercase; color: var(--wd-fg); font-weight: 500;
-}
-.wd-stat { display: inline-flex; align-items: center; gap: 5px; font-size: 12px; font-weight: 600; }
-.wd-stat svg { width: 13px; height: 13px; flex: none; }
-.wd-covered { color: var(--wd-ok); }
-.wd-partial { color: var(--wd-warn); }
-.wd-absent { color: var(--wd-danger); }
-.wd-pin { font-family: var(--wd-mono); font-size: 10px; color: var(--wd-fg-4); }
-
-.wd-status { font-family: var(--wd-mono); font-size: 11px; flex: none; }
-.wd-status-running { color: var(--wd-warn); }
-.wd-status-done { color: var(--wd-ok); }
-.wd-status-failed { color: var(--wd-danger); }
-.wd-output {
-  flex-basis: 100%;
-  width: 100%;
-  font-family: var(--wd-mono); font-size: 11px; white-space: pre-wrap;
-  max-height: 220px; overflow: auto; margin-top: 8px;
-  border: 1px solid var(--wd-border); border-radius: 4px; padding: 8px;
-  color: var(--wd-fg-2);
-}
+.webinar-dash {
+  --wd-ink-000: #ffffff; --wd-ink-050: #f7f7f7; --wd-ink-100: #ececec;
+  --wd-ink-200: #d9d9d9; --wd-ink-400: #8a8a8a; --wd-ink-500: #5e5e5e;
+  --wd-ink-700: #262626; --wd-ink-900: #0a0a0a; --wd-ink-999: #000000;
+  --wd-red-500: #e1261c; --wd-red-600: #c31c14;
+
+  --wd-bg: var(--wd-ink-000);
+  --wd-bg-subtle: var(--wd-ink-050);
+  --wd-fg: var(--wd-ink-999);
+  --wd-fg-2: var(--wd-ink-700);
+  --wd-fg-3: var(--wd-ink-500);
+  --wd-fg-4: var(--wd-ink-400);
+  --wd-border: var(--wd-ink-200);
+  --wd-border-strong: var(--wd-ink-999);
+  --wd-border-subtle: var(--wd-ink-100);
+  --wd-accent: var(--wd-red-500);
+  --wd-accent-press: var(--wd-red-600);
+  --wd-accent-on: #ffffff;
+  --wd-ok: #0a8a3f;
+  --wd-warn: #c68a00;
+  --wd-danger: var(--wd-red-500);
+
+  --wd-mono: "JetBrains Mono", ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
+  --wd-sans: "Inter", ui-sans-serif, system-ui, -apple-system, "Helvetica Neue", Arial, sans-serif;
+
+  font-family: var(--wd-sans);
+  color: var(--wd-fg);
+  display: flex;
+  flex-direction: column;
+  gap: 16px;
+  /* Past roughly this width the two panes stop reading as a pair and table
+     rows get hard to track across. Cap the whole dashboard, not just the
+     grid, so the toolbar stays flush with the panes below it. */
+  max-width: 1600px;
+  /* Makes this element the reference for the @container query further down,
+     so the grid collapses on the pane's width rather than the window's. */
+  container-type: inline-size;
+}
+
+.theme-dark .webinar-dash {
+  --wd-bg: var(--wd-ink-900);
+  --wd-bg-subtle: #161616;
+  --wd-fg: var(--wd-ink-000);
+  --wd-fg-2: var(--wd-ink-200);
+  --wd-fg-3: var(--wd-ink-400);
+  --wd-fg-4: var(--wd-ink-500);
+  --wd-border: var(--wd-ink-700);
+  --wd-border-strong: #b8b8b8;
+  --wd-border-subtle: #161616;
+  /* Accent tracks the installed theme at .obsidian/themes/tesanti/theme.css,
+     whose dark section reads "Black canvas, same signal red": --accent-h/-s/-l
+     are declared once at :root and never redeclared under .theme-dark, and only
+     the hover state lifts (#c31c14 light, #ff4d43 dark). Match that exactly. */
+  --wd-accent: var(--wd-red-500);
+  --wd-accent-press: #ff4d43;
+  --wd-accent-on: #ffffff;
+  /* Status tokens are separate from the brand accent by design-system rule, and
+     here they carry small text in a dense table. #e1261c on near-black is about
+     3.8:1, under AA for small text, so these lift where the accent does not. */
+  --wd-ok: #2fbf6a;
+  --wd-warn: #e0a516;
+  --wd-danger: #ff5c50;
+}
+
+/* ------------------------------------------------------------------
+   Escape Obsidian's readable-line-length cap.
+
+   With "Readable line length" on (the default), Obsidian caps note
+   content at --file-line-width, roughly 700px. That is right for prose
+   and wrong for a two-pane dashboard, which gets squeezed into a column.
+
+   :has() scopes this to the sizer that actually contains a dashboard, so
+   every other note in the vault keeps its readable width. Both selectors
+   are needed: reading view sizes on .markdown-preview-sizer, live preview
+   on .cm-sizer.
+
+   If you would rather not rely on this, the alternatives are Settings ->
+   Editor -> Readable line length (off, but that widens every note), or
+   adding `cssclasses: wide-dash` to the note's frontmatter and swapping
+   the :has() selectors below for `.wide-dash .markdown-preview-sizer`.
+   ------------------------------------------------------------------ */
+.markdown-preview-sizer:has(.webinar-dash),
+.markdown-source-view.mod-cm6 .cm-sizer:has(.webinar-dash) {
+  max-width: none;
+}
+
+.wd-grid {
+  display: grid;
+  grid-template-columns: minmax(0, 38fr) minmax(0, 62fr);
+  gap: 24px;
+}
+.wd-grid > * { min-width: 0; }
+
+.wd-bar {
+  display: flex;
+  align-items: center;
+  gap: 12px;
+  flex-wrap: wrap;
+  border: 1px solid var(--wd-border);
+  border-radius: 4px;
+  padding: 8px 12px;
+  background: var(--wd-bg-subtle);
+}
+
+.wd-btn-ghost {
+  background: transparent;
+  color: var(--wd-fg);
+  border-color: var(--wd-border-strong);
+}
+.wd-btn-ghost:hover {
+  background: var(--wd-bg-muted, var(--wd-bg-subtle));
+  border-color: var(--wd-border-strong);
+}
+
+/* Collapse to one column on the width that actually matters — the pane's,
+   not the window's. A media query measures the window, so a wide window
+   with the dashboard in a narrow split pane would keep two cramped
+   columns. The container query below responds to the pane itself; the
+   media query stays as a floor for a genuinely narrow window. */
+@container (max-width: 820px) {
+  .wd-grid { grid-template-columns: minmax(0, 1fr); }
+}
+
+@media (max-width: 820px) {
+  .wd-grid { grid-template-columns: minmax(0, 1fr); }
+}
+
+.wd-pane { display: flex; flex-direction: column; gap: 20px; }
+
+.wd-eyebrow {
+  font-family: var(--wd-mono); font-size: 11px; line-height: 1;
+  letter-spacing: 0.12em; text-transform: uppercase;
+  color: var(--wd-fg-3); font-weight: 500;
+  display: flex; align-items: center; gap: 8px;
+}
+.wd-eyebrow::before {
+  content: ""; width: 5px; height: 5px; flex: none; background: var(--wd-accent);
+}
+
+.wd-block { display: flex; flex-direction: column; gap: 12px; }
+
+.wd-row {
+  display: flex; align-items: center; gap: 12px; flex-wrap: wrap;
+  border: 1px solid var(--wd-border); border-radius: 4px;
+  padding: 12px 12px 12px 16px; background: var(--wd-bg);
+}
+.wd-row + .wd-row { margin-top: -1px; }
+.wd-row-main { flex: 1; min-width: 0; display: flex; flex-direction: column; gap: 2px; }
+.wd-row-name {
+  font-size: 14px; font-weight: 600;
+  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
+}
+.wd-mono {
+  font-family: var(--wd-mono); font-size: 11px; color: var(--wd-fg-3);
+  font-variant-numeric: tabular-nums;
+}
+
+.wd-btn {
+  display: inline-flex; align-items: center; gap: 6px; flex: none;
+  font-family: var(--wd-sans); font-size: 13px; font-weight: 600;
+  padding: 6px 12px; border-radius: 6px; cursor: pointer;
+  border: 1px solid var(--wd-accent); background: var(--wd-accent);
+  color: var(--wd-accent-on);
+  transition: background 120ms cubic-bezier(0.2, 0, 0, 1);
+}
+.wd-btn:hover { background: var(--wd-accent-press); border-color: var(--wd-accent-press); }
+.wd-btn:focus-visible { outline: 2px solid var(--wd-accent); outline-offset: 2px; }
+.wd-btn[disabled] {
+  opacity: 0.45; cursor: not-allowed;
+  background: transparent; color: var(--wd-fg-3); border-color: var(--wd-border);
+}
+.wd-btn svg { width: 14px; height: 14px; flex: none; }
+
+.wd-done { display: flex; flex-direction: column; }
+.wd-done-row {
+  display: flex; align-items: baseline; gap: 12px; padding: 5px 0;
+  border-bottom: 1px solid var(--wd-border-subtle); font-size: 13px;
+}
+.wd-done-row:last-child { border-bottom: 0; }
+.wd-done-name {
+  color: var(--wd-fg-2);
+  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
+}
+
+.wd-note { font-size: 12px; color: var(--wd-fg-3); }
+.wd-error {
+  font-size: 13px; color: var(--wd-danger);
+  border: 1px solid var(--wd-danger); border-radius: 4px; padding: 12px;
+}
+
+@media (prefers-reduced-motion: reduce) {
+  .webinar-dash * { transition-duration: 0.01ms !important; }
+}
+
+.wd-meter { display: flex; gap: 2px; height: 8px; width: 100%; }
+.wd-meter > i { display: block; height: 100%; }
+.wd-seg-covered { background: var(--wd-ok); }
+.wd-seg-partial { background: var(--wd-warn); }
+.wd-seg-absent { background: var(--wd-danger); }
+
+.wd-key { display: flex; flex-wrap: wrap; gap: 16px; font-size: 12px; color: var(--wd-fg-2); }
+.wd-key > span { display: inline-flex; align-items: center; gap: 6px; }
+.wd-key i { width: 8px; height: 8px; flex: none; }
+
+.wd-tbl { width: 100%; border-collapse: collapse; font-size: 13px; }
+.wd-tbl th {
+  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
+  text-transform: uppercase; color: var(--wd-fg-3); font-weight: 500;
+  text-align: left; padding: 0 12px 8px 0;
+  border-bottom: 1px solid var(--wd-border-strong);
+}
+.wd-tbl td {
+  padding: 6px 12px 6px 0; border-bottom: 1px solid var(--wd-border-subtle);
+  vertical-align: baseline;
+}
+.wd-grp td {
+  padding-top: 16px; border-bottom: 1px solid var(--wd-border-strong);
+  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
+  text-transform: uppercase; color: var(--wd-fg); font-weight: 500;
+}
+.wd-stat { display: inline-flex; align-items: center; gap: 5px; font-size: 12px; font-weight: 600; }
+.wd-stat svg { width: 13px; height: 13px; flex: none; }
+.wd-covered { color: var(--wd-ok); }
+.wd-partial { color: var(--wd-warn); }
+.wd-absent { color: var(--wd-danger); }
+.wd-pin { font-family: var(--wd-mono); font-size: 10px; color: var(--wd-fg-4); }
+
+.wd-status { font-family: var(--wd-mono); font-size: 11px; flex: none; }
+.wd-status-running { color: var(--wd-warn); }
+.wd-status-done { color: var(--wd-ok); }
+.wd-status-failed { color: var(--wd-danger); }
+.wd-output {
+  flex-basis: 100%;
+  width: 100%;
+  font-family: var(--wd-mono); font-size: 11px; white-space: pre-wrap;
+  max-height: 220px; overflow: auto; margin-top: 8px;
+  border: 1px solid var(--wd-border); border-radius: 4px; padding: 8px;
+  color: var(--wd-fg-2);
+}
diff --git a/.obsidian/plugins/webinar-dash/test/coverage.test.js b/.obsidian/plugins/webinar-dash/test/coverage.test.js
index 22ff85f..c5f57f2 100644
--- a/.obsidian/plugins/webinar-dash/test/coverage.test.js
+++ b/.obsidian/plugins/webinar-dash/test/coverage.test.js
@@ -1,122 +1,122 @@
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { parseCoverageTable, groupByStation } = require("../main.js").__test__;
-
-const DOC = [
-  "# Script coverage",
-  "",
-  "#coverage",
-  "",
-  "## Metadata",
-  "",
-  "- **Script:** `raw/sources/Webinar script.md`",
-  "- **Last synced:** 2026-07-28",
-  "",
-  "## Coverage",
-  "",
-  "| Concept | Status | Station | Pinned |",
-  "|---|---|---|---|",
-  "| [[harness]] | covered | Tools | |",
-  "| [[agentic-loops]] | partial | Process | |",
-  "| [[levels-of-ai-usage]] | partial | all | |",
-  "| [[connections-as-moat]] | absent | — | yes |",
-].join("\n");
-
-test("parseCoverageTable reads metadata", () => {
-  const { meta } = parseCoverageTable(DOC);
-  assert.equal(meta.script, "raw/sources/Webinar script.md");
-  assert.equal(meta.lastSynced, "2026-07-28");
-});
-
-test("parseCoverageTable reads every data row and skips the header", () => {
-  const { rows, errors } = parseCoverageTable(DOC);
-  assert.equal(errors.length, 0);
-  assert.equal(rows.length, 4);
-  assert.deepEqual(rows.map((r) => r.concept), [
-    "harness", "agentic-loops", "levels-of-ai-usage", "connections-as-moat",
-  ]);
-});
-
-test("parseCoverageTable normalises stations", () => {
-  const { rows } = parseCoverageTable(DOC);
-  assert.deepEqual(rows[0].stations, ["Tools"]);
-  assert.deepEqual(rows[2].stations, ["all"]);
-  assert.deepEqual(rows[3].stations, []);
-});
-
-test("parseCoverageTable reads the pinned flag", () => {
-  const { rows } = parseCoverageTable(DOC);
-  assert.equal(rows[0].pinned, false);
-  assert.equal(rows[3].pinned, true);
-});
-
-test("parseCoverageTable strips a wikilink alias", () => {
-  const doc = "| Concept | Status | Station |\n|---|---|---|\n| [[harness\\|The harness]] | covered | Tools |";
-  const { rows } = parseCoverageTable(doc);
-  assert.equal(rows[0].concept, "harness");
-});
-
-test("parseCoverageTable splits a multi-station cell", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered | Tools, Memory |";
-  const { rows } = parseCoverageTable(doc);
-  assert.deepEqual(rows[0].stations, ["Tools", "Memory"]);
-});
-
-test("parseCoverageTable rejects an invalid status instead of coercing it", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | maybe | Tools |";
-  const { rows, errors } = parseCoverageTable(doc);
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 1);
-  assert.match(errors[0].reason, /invalid status/);
-  assert.equal(errors[0].line, 3);
-});
-
-test("parseCoverageTable reports a row with too few columns", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered |";
-  const { rows, errors } = parseCoverageTable(doc);
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 1);
-  assert.match(errors[0].reason, /at least 3 columns/);
-});
-
-test("parseCoverageTable returns empty results for a document with no table", () => {
-  const { rows, errors } = parseCoverageTable("# Nothing\n\nJust prose.");
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 0);
-});
-
-test("groupByStation orders all-stations first and no-station last", () => {
-  const { rows } = parseCoverageTable(DOC);
-  const groups = groupByStation(rows);
-  assert.deepEqual(groups.map((g) => g.station), [
-    "All stations", "Tools", "Process", "No station",
-  ]);
-  assert.equal(groups[1].rows[0].concept, "harness");
-});
-
-test("groupByStation places a multi-station row under its first station only", () => {
-  const rows = [{ concept: "x", status: "covered", stations: ["Memory", "Skills"], pinned: false, line: 1 }];
-  const groups = groupByStation(rows);
-  assert.equal(groups.length, 1);
-  assert.equal(groups[0].station, "Memory");
-});
-
-const { reconcileConcepts } = require("../main.js").__test__;
-
-test("reconcileConcepts finds concept pages with no row", () => {
-  const rows = [{ concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 }];
-  const out = reconcileConcepts(rows, ["harness", "brand-new-concept"]);
-  assert.deepEqual(out.unsynced, ["brand-new-concept"]);
-  assert.deepEqual(out.stale, []);
-});
-
-test("reconcileConcepts finds rows whose concept page is gone", () => {
-  const rows = [
-    { concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 },
-    { concept: "deleted-idea", status: "absent", stations: [], pinned: false, line: 2 },
-  ];
-  const out = reconcileConcepts(rows, ["harness"]);
-  assert.deepEqual(out.stale.map((r) => r.concept), ["deleted-idea"]);
-  assert.deepEqual(out.unsynced, []);
-});
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { parseCoverageTable, groupByStation } = require("../main.js").__test__;
+
+const DOC = [
+  "# Script coverage",
+  "",
+  "#coverage",
+  "",
+  "## Metadata",
+  "",
+  "- **Script:** `raw/sources/Webinar script.md`",
+  "- **Last synced:** 2026-07-28",
+  "",
+  "## Coverage",
+  "",
+  "| Concept | Status | Station | Pinned |",
+  "|---|---|---|---|",
+  "| [[harness]] | covered | Tools | |",
+  "| [[agentic-loops]] | partial | Process | |",
+  "| [[levels-of-ai-usage]] | partial | all | |",
+  "| [[connections-as-moat]] | absent | — | yes |",
+].join("\n");
+
+test("parseCoverageTable reads metadata", () => {
+  const { meta } = parseCoverageTable(DOC);
+  assert.equal(meta.script, "raw/sources/Webinar script.md");
+  assert.equal(meta.lastSynced, "2026-07-28");
+});
+
+test("parseCoverageTable reads every data row and skips the header", () => {
+  const { rows, errors } = parseCoverageTable(DOC);
+  assert.equal(errors.length, 0);
+  assert.equal(rows.length, 4);
+  assert.deepEqual(rows.map((r) => r.concept), [
+    "harness", "agentic-loops", "levels-of-ai-usage", "connections-as-moat",
+  ]);
+});
+
+test("parseCoverageTable normalises stations", () => {
+  const { rows } = parseCoverageTable(DOC);
+  assert.deepEqual(rows[0].stations, ["Tools"]);
+  assert.deepEqual(rows[2].stations, ["all"]);
+  assert.deepEqual(rows[3].stations, []);
+});
+
+test("parseCoverageTable reads the pinned flag", () => {
+  const { rows } = parseCoverageTable(DOC);
+  assert.equal(rows[0].pinned, false);
+  assert.equal(rows[3].pinned, true);
+});
+
+test("parseCoverageTable strips a wikilink alias", () => {
+  const doc = "| Concept | Status | Station |\n|---|---|---|\n| [[harness\\|The harness]] | covered | Tools |";
+  const { rows } = parseCoverageTable(doc);
+  assert.equal(rows[0].concept, "harness");
+});
+
+test("parseCoverageTable splits a multi-station cell", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered | Tools, Memory |";
+  const { rows } = parseCoverageTable(doc);
+  assert.deepEqual(rows[0].stations, ["Tools", "Memory"]);
+});
+
+test("parseCoverageTable rejects an invalid status instead of coercing it", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | maybe | Tools |";
+  const { rows, errors } = parseCoverageTable(doc);
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 1);
+  assert.match(errors[0].reason, /invalid status/);
+  assert.equal(errors[0].line, 3);
+});
+
+test("parseCoverageTable reports a row with too few columns", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered |";
+  const { rows, errors } = parseCoverageTable(doc);
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 1);
+  assert.match(errors[0].reason, /at least 3 columns/);
+});
+
+test("parseCoverageTable returns empty results for a document with no table", () => {
+  const { rows, errors } = parseCoverageTable("# Nothing\n\nJust prose.");
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 0);
+});
+
+test("groupByStation orders all-stations first and no-station last", () => {
+  const { rows } = parseCoverageTable(DOC);
+  const groups = groupByStation(rows);
+  assert.deepEqual(groups.map((g) => g.station), [
+    "All stations", "Tools", "Process", "No station",
+  ]);
+  assert.equal(groups[1].rows[0].concept, "harness");
+});
+
+test("groupByStation places a multi-station row under its first station only", () => {
+  const rows = [{ concept: "x", status: "covered", stations: ["Memory", "Skills"], pinned: false, line: 1 }];
+  const groups = groupByStation(rows);
+  assert.equal(groups.length, 1);
+  assert.equal(groups[0].station, "Memory");
+});
+
+const { reconcileConcepts } = require("../main.js").__test__;
+
+test("reconcileConcepts finds concept pages with no row", () => {
+  const rows = [{ concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 }];
+  const out = reconcileConcepts(rows, ["harness", "brand-new-concept"]);
+  assert.deepEqual(out.unsynced, ["brand-new-concept"]);
+  assert.deepEqual(out.stale, []);
+});
+
+test("reconcileConcepts finds rows whose concept page is gone", () => {
+  const rows = [
+    { concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 },
+    { concept: "deleted-idea", status: "absent", stations: [], pinned: false, line: 2 },
+  ];
+  const out = reconcileConcepts(rows, ["harness"]);
+  assert.deepEqual(out.stale.map((r) => r.concept), ["deleted-idea"]);
+  assert.deepEqual(out.unsynced, []);
+});
diff --git a/.obsidian/plugins/webinar-dash/test/pipeline.test.js b/.obsidian/plugins/webinar-dash/test/pipeline.test.js
index 61b39cb..b1f86cf 100644
--- a/.obsidian/plugins/webinar-dash/test/pipeline.test.js
+++ b/.obsidian/plugins/webinar-dash/test/pipeline.test.js
@@ -1,109 +1,109 @@
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { extractRawPath, derivePipeline } = require("../main.js").__test__;
-
-test("extractRawPath pulls the backticked path", () => {
-  const page = [
-    "# You're reading way too much code",
-    "",
-    "#source",
-    "",
-    "## Source Metadata",
-    "",
-    "- **Date:** YouTube video, 24:11",
-    "- **Raw path:** `raw/sources/You're reading way too much code.md`",
-    "- **Source type:** video essay",
-  ].join("\n");
-  assert.equal(extractRawPath(page), "raw/sources/You're reading way too much code.md");
-});
-
-test("extractRawPath handles cyrillic and em dashes", () => {
-  const page = "- **Raw path:** `raw/sources/Скиллы на базе git — новая память AI-агентов.md`";
-  assert.equal(extractRawPath(page), "raw/sources/Скиллы на базе git — новая память AI-агентов.md");
-});
-
-test("extractRawPath returns null when the line is absent", () => {
-  assert.equal(extractRawPath("# A page\n\n#source\n\nNo metadata here."), null);
-});
-
-test("derivePipeline splits claimed from unclaimed raw files", () => {
-  const rawFiles = [
-    { path: "raw/sources/Nina interview.md", name: "Nina interview.md", size: 6614 },
-    { path: "raw/sources/Webinar script.md", name: "Webinar script.md", size: 15841 },
-  ];
-  const sourcePages = [
-    {
-      path: "wiki/sources/2026-07-14-nina-interview.md",
-      name: "2026-07-14-nina-interview.md",
-      rawPath: "raw/sources/Nina interview.md",
-    },
-  ];
-  const out = derivePipeline({ rawFiles, sourcePages });
-  assert.equal(out.processed.length, 1);
-  assert.equal(out.processed[0].name, "Nina interview.md");
-  assert.equal(out.processed[0].page.name, "2026-07-14-nina-interview.md");
-  assert.equal(out.unprocessed.length, 1);
-  assert.equal(out.unprocessed[0].name, "Webinar script.md");
-  assert.equal(out.orphaned.length, 0);
-});
-
-test("derivePipeline reports source pages whose raw file is gone", () => {
-  const out = derivePipeline({
-    rawFiles: [],
-    sourcePages: [
-      { path: "wiki/sources/x.md", name: "x.md", rawPath: "raw/sources/deleted.md" },
-    ],
-  });
-  assert.equal(out.orphaned.length, 1);
-  assert.equal(out.orphaned[0].name, "x.md");
-});
-
-test("derivePipeline treats a page with no raw path as orphaned", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
-    sourcePages: [{ path: "wiki/sources/y.md", name: "y.md", rawPath: null }],
-  });
-  assert.equal(out.orphaned.length, 1);
-  assert.equal(out.unprocessed.length, 1);
-});
-
-test("derivePipeline routes a duplicate raw-path claim to orphaned", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
-    sourcePages: [
-      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/a.md" },
-      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/a.md" },
-    ],
-  });
-  assert.equal(out.processed.length, 1);
-  assert.equal(out.processed[0].page.name, "first.md");
-  assert.equal(out.unprocessed.length, 0);
-  assert.deepEqual(out.orphaned.map((p) => p.name), ["second.md"]);
-});
-
-test("derivePipeline lists a page once when it is both a duplicate claim and missing its raw file", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/other.md", name: "other.md", size: 10 }],
-    sourcePages: [
-      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/gone.md" },
-      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/gone.md" },
-    ],
-  });
-  assert.deepEqual(out.orphaned.map((p) => p.name), ["first.md", "second.md"]);
-});
-
-test("derivePipeline sorts unprocessed by name and processed newest first", () => {
-  const out = derivePipeline({
-    rawFiles: [
-      { path: "raw/sources/b.md", name: "b.md", size: 1 },
-      { path: "raw/sources/a.md", name: "a.md", size: 1 },
-      { path: "raw/sources/c.md", name: "c.md", size: 1 },
-    ],
-    sourcePages: [
-      { path: "wiki/sources/2026-07-14-x.md", name: "2026-07-14-x.md", rawPath: "raw/sources/c.md" },
-    ],
-  });
-  assert.deepEqual(out.unprocessed.map((f) => f.name), ["a.md", "b.md"]);
-  assert.equal(out.processed.length, 1);
-});
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { extractRawPath, derivePipeline } = require("../main.js").__test__;
+
+test("extractRawPath pulls the backticked path", () => {
+  const page = [
+    "# You're reading way too much code",
+    "",
+    "#source",
+    "",
+    "## Source Metadata",
+    "",
+    "- **Date:** YouTube video, 24:11",
+    "- **Raw path:** `raw/sources/You're reading way too much code.md`",
+    "- **Source type:** video essay",
+  ].join("\n");
+  assert.equal(extractRawPath(page), "raw/sources/You're reading way too much code.md");
+});
+
+test("extractRawPath handles cyrillic and em dashes", () => {
+  const page = "- **Raw path:** `raw/sources/Скиллы на базе git — новая память AI-агентов.md`";
+  assert.equal(extractRawPath(page), "raw/sources/Скиллы на базе git — новая память AI-агентов.md");
+});
+
+test("extractRawPath returns null when the line is absent", () => {
+  assert.equal(extractRawPath("# A page\n\n#source\n\nNo metadata here."), null);
+});
+
+test("derivePipeline splits claimed from unclaimed raw files", () => {
+  const rawFiles = [
+    { path: "raw/sources/Nina interview.md", name: "Nina interview.md", size: 6614 },
+    { path: "raw/sources/Webinar script.md", name: "Webinar script.md", size: 15841 },
+  ];
+  const sourcePages = [
+    {
+      path: "wiki/sources/2026-07-14-nina-interview.md",
+      name: "2026-07-14-nina-interview.md",
+      rawPath: "raw/sources/Nina interview.md",
+    },
+  ];
+  const out = derivePipeline({ rawFiles, sourcePages });
+  assert.equal(out.processed.length, 1);
+  assert.equal(out.processed[0].name, "Nina interview.md");
+  assert.equal(out.processed[0].page.name, "2026-07-14-nina-interview.md");
+  assert.equal(out.unprocessed.length, 1);
+  assert.equal(out.unprocessed[0].name, "Webinar script.md");
+  assert.equal(out.orphaned.length, 0);
+});
+
+test("derivePipeline reports source pages whose raw file is gone", () => {
+  const out = derivePipeline({
+    rawFiles: [],
+    sourcePages: [
+      { path: "wiki/sources/x.md", name: "x.md", rawPath: "raw/sources/deleted.md" },
+    ],
+  });
+  assert.equal(out.orphaned.length, 1);
+  assert.equal(out.orphaned[0].name, "x.md");
+});
+
+test("derivePipeline treats a page with no raw path as orphaned", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
+    sourcePages: [{ path: "wiki/sources/y.md", name: "y.md", rawPath: null }],
+  });
+  assert.equal(out.orphaned.length, 1);
+  assert.equal(out.unprocessed.length, 1);
+});
+
+test("derivePipeline routes a duplicate raw-path claim to orphaned", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
+    sourcePages: [
+      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/a.md" },
+      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/a.md" },
+    ],
+  });
+  assert.equal(out.processed.length, 1);
+  assert.equal(out.processed[0].page.name, "first.md");
+  assert.equal(out.unprocessed.length, 0);
+  assert.deepEqual(out.orphaned.map((p) => p.name), ["second.md"]);
+});
+
+test("derivePipeline lists a page once when it is both a duplicate claim and missing its raw file", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/other.md", name: "other.md", size: 10 }],
+    sourcePages: [
+      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/gone.md" },
+      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/gone.md" },
+    ],
+  });
+  assert.deepEqual(out.orphaned.map((p) => p.name), ["first.md", "second.md"]);
+});
+
+test("derivePipeline sorts unprocessed by name and processed newest first", () => {
+  const out = derivePipeline({
+    rawFiles: [
+      { path: "raw/sources/b.md", name: "b.md", size: 1 },
+      { path: "raw/sources/a.md", name: "a.md", size: 1 },
+      { path: "raw/sources/c.md", name: "c.md", size: 1 },
+    ],
+    sourcePages: [
+      { path: "wiki/sources/2026-07-14-x.md", name: "2026-07-14-x.md", rawPath: "raw/sources/c.md" },
+    ],
+  });
+  assert.deepEqual(out.unprocessed.map((f) => f.name), ["a.md", "b.md"]);
+  assert.equal(out.processed.length, 1);
+});
diff --git a/.obsidian/plugins/webinar-dash/test/safety.test.js b/.obsidian/plugins/webinar-dash/test/safety.test.js
index 419dc1b..a9eb557 100644
--- a/.obsidian/plugins/webinar-dash/test/safety.test.js
+++ b/.obsidian/plugins/webinar-dash/test/safety.test.js
@@ -1,47 +1,47 @@
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { isSafeFilename } = require("../main.js").__test__;
-
-test("isSafeFilename accepts every filename currently in the vault", () => {
-  const real = [
-    "Agentic Engineering, explained by a 10x developer.md",
-    "Webinar Plan - From Chat Box to Your Own OS.md",
-    "Webinar script.md",
-    "You're reading way too much code.md",
-    "ИИ глупый!.md",
-    "Скиллы на базе git — новая память AI-агентов.md",
-    "sebastian interview - conclusions and insights.md",
-    "In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md",
-  ];
-  for (const name of real) {
-    assert.equal(isSafeFilename(name), true, `should accept: ${name}`);
-  }
-});
-
-test("isSafeFilename rejects shell metacharacters", () => {
-  for (const bad of ['a".md', "a`b.md", "a$b.md", "a&b.md", "a|b.md", "a;b.md",
-                     "a<b.md", "a>b.md", "a%b.md", "a\nb.md", "a\rb.md"]) {
-    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
-  }
-});
-
-test("isSafeFilename rejects the cmd.exe escape character and control characters", () => {
-  // `^` escapes the next character in cmd.exe, so it can defuse the closing
-  // quote. NUL additionally makes spawn() throw synchronously.
-  for (const bad of ["a^b.md", "a\u0000b.md", "a\u001bb.md", "a\u007fb.md"]) {
-    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
-  }
-});
-
-test("isSafeFilename rejects path traversal", () => {
-  assert.equal(isSafeFilename("../secrets.md"), false);
-  assert.equal(isSafeFilename("a/../../b.md"), false);
-});
-
-test("isSafeFilename rejects empty and non-string input", () => {
-  assert.equal(isSafeFilename(""), false);
-  assert.equal(isSafeFilename(null), false);
-  assert.equal(isSafeFilename(undefined), false);
-  assert.equal(isSafeFilename(42), false);
-});
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { isSafeFilename } = require("../main.js").__test__;
+
+test("isSafeFilename accepts every filename currently in the vault", () => {
+  const real = [
+    "Agentic Engineering, explained by a 10x developer.md",
+    "Webinar Plan - From Chat Box to Your Own OS.md",
+    "Webinar script.md",
+    "You're reading way too much code.md",
+    "ИИ глупый!.md",
+    "Скиллы на базе git — новая память AI-агентов.md",
+    "sebastian interview - conclusions and insights.md",
+    "In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md",
+  ];
+  for (const name of real) {
+    assert.equal(isSafeFilename(name), true, `should accept: ${name}`);
+  }
+});
+
+test("isSafeFilename rejects shell metacharacters", () => {
+  for (const bad of ['a".md', "a`b.md", "a$b.md", "a&b.md", "a|b.md", "a;b.md",
+                     "a<b.md", "a>b.md", "a%b.md", "a\nb.md", "a\rb.md"]) {
+    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
+  }
+});
+
+test("isSafeFilename rejects the cmd.exe escape character and control characters", () => {
+  // `^` escapes the next character in cmd.exe, so it can defuse the closing
+  // quote. NUL additionally makes spawn() throw synchronously.
+  for (const bad of ["a^b.md", "a\u0000b.md", "a\u001bb.md", "a\u007fb.md"]) {
+    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
+  }
+});
+
+test("isSafeFilename rejects path traversal", () => {
+  assert.equal(isSafeFilename("../secrets.md"), false);
+  assert.equal(isSafeFilename("a/../../b.md"), false);
+});
+
+test("isSafeFilename rejects empty and non-string input", () => {
+  assert.equal(isSafeFilename(""), false);
+  assert.equal(isSafeFilename(null), false);
+  assert.equal(isSafeFilename(undefined), false);
+  assert.equal(isSafeFilename(42), false);
+});
diff --git a/CLAUDE.md b/CLAUDE.md
index fc27eef..ac3114e 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -1,248 +1,248 @@
-# LLM Wiki Schema (Operational Contract)
-
-You are the dedicated maintainer of this vault as a persistent, compounding wiki.
-
-## Mission
-
-Maintain a high-signal personal knowledge base where:
-
-- `raw/` contains immutable source materials.
-- `wiki/` contains LLM-authored, continuously maintained markdown pages.
-- `index.md` is the content catalog.
-- `log.md` is an append-only chronological operation log.
-
-The user curates sources and asks questions. You perform all wiki maintenance.
-
-## Non-Negotiable Rules
-
-1. Never modify files inside `raw/`.
-2. Always update `index.md` and `log.md` after any ingest/query/lint operation that changes the wiki.
-3. Prefer editing existing pages over creating duplicates.
-4. Use Obsidian-style wiki links (`[[page-name]]`) for internal references.
-5. Record contradictions and superseded claims explicitly (do not silently overwrite history).
-6. Keep claims attributable: include source links to `wiki/sources/*` pages.
-7. Do not leave orphan pages intentionally; add at least one inbound and one outbound link when possible.
-8. Keep writing concise, structured, and diff-friendly.
-9. Every wiki page carries exactly one page-type hashtag (see Tagging Rules) matching its folder.
-
-## Folder Convention
-
-```text
-raw/
-  sources/            # immutable source markdown/text/pdf exports
-  assets/             # immutable local images/files referenced by raw sources
-
-wiki/
-  overview.md         # top-level synthesis and navigation
-  script-coverage.md  # machine-maintained: concept coverage vs the webinar script
-  sources/            # one summary page per ingested raw source
-  entities/           # people, orgs, projects, places, tools
-  concepts/           # themes, ideas, methods, frameworks
-  timelines/          # optional chronological reconstructions
-  comparisons/        # side-by-side analyses
-  queries/            # durable outputs created from Q&A sessions
-  lint-reports/       # periodic health-check reports
-```
-
-## File Naming Rules
-
-- Use kebab-case for file names.
-- Prefix source summary pages with date: `YYYY-MM-DD-title.md`.
-- Prefer stable canonical pages:
-  - `wiki/entities/<name>.md`
-  - `wiki/concepts/<concept>.md`
-- If a page name collides, merge instead of creating suffixes unless truly distinct.
-
-## Tagging Rules
-
-Every page in `wiki/` starts with its page-type hashtag on its own line, directly after the `# <title>` heading (Obsidian inline tag format):
-
-```text
-# <Title>
-
-#<type-tag>
-
-## <first section>
-```
-
-Folder → required tag:
-
-| Folder / file          | Tag            |
-| ---------------------- | -------------- |
-| `wiki/overview.md`     | `#overview`    |
-| `wiki/script-coverage.md` | `#coverage`    |
-| `wiki/sources/*`       | `#source`      |
-| `wiki/entities/*`      | `#entity`      |
-| `wiki/concepts/*`      | `#concept`     |
-| `wiki/timelines/*`     | `#timeline`    |
-| `wiki/comparisons/*`   | `#comparison`  |
-| `wiki/queries/*`       | `#query`       |
-| `wiki/lint-reports/*`  | `#lint-report` |
-
-- Exactly one page-type tag per page; it must match the page's folder.
-- Additional topical tags are allowed on the same line after the type tag (e.g. `#concept #ai-tooling`), but the type tag comes first.
-- When creating any new page, add the tag immediately; when moving a page between folders, update the tag in the same operation.
-
-## Required Page Templates
-
-### 1) Source Summary (`wiki/sources/*.md`)
-
-Must include sections:
-
-- `# <title>`
-- `#source` tag line (per Tagging Rules)
-- `## Source Metadata` (date, raw path, source type, ingestion date)
-- `## Core Claims`
-- `## Key Evidence / Details`
-- `## Connections` (links to entities/concepts/comparisons)
-- `## Open Questions`
-- `## Change Impact on Wiki` (what pages were updated and why)
-
-### 2) Entity / Concept Page
-
-Must include:
-
-- `# <name>`
-- `#entity` or `#concept` tag line (per Tagging Rules)
-- `## Summary`
-- `## Current Understanding`
-- `## Evidence` (bullets with links to `wiki/sources/*`)
-- `## Related Pages`
-- `## Contradictions / Uncertainty`
-- `## Next Questions`
-
-### 3) Query Output (`wiki/queries/*.md`)
-
-Must include:
-
-- `#query` tag line (per Tagging Rules)
-- Question asked
-- Answer
-- Evidence trail (links)
-- Follow-up questions
-- Whether this output changed existing pages
-
-### 4) Coverage File (`wiki/script-coverage.md`)
-
-Machine-read by the dashboard plugin, which parses the table positionally. The
-shape below is required exactly — a deviation renders as parse errors, not as a
-best-effort read.
-
-- `# Script coverage`
-- `#coverage` tag line
-- `## Metadata`, containing these two lines verbatim in this form:
-  - `- **Script:** ` followed by the raw script path in backticks
-  - `- **Last synced:** ` followed by a `YYYY-MM-DD` date
-- `## Coverage`, containing one markdown table with exactly these four columns in
-  this order: `Concept`, `Status`, `Station`, `Pinned`
-  - **Concept** — an Obsidian wikilink to a page in `wiki/concepts/`
-  - **Status** — exactly one of `covered`, `partial`, `absent`, lowercase
-  - **Station** — one of the seven station names spelled exactly as listed in
-    Workflow D, a comma-separated list of them, the lowercase word `all`, or an
-    em dash `—` when the status is `absent`
-  - **Pinned** — the lowercase word `yes`, or empty
-- `## Notes` and `## Related Pages` are free prose and are not parsed.
-
-The parser reads every markdown table in the file, so do not add a second table.
-
-## Standard Workflows
-
-### Workflow A: Ingest One Source
-
-When user says "ingest <source>":
-
-1. Read the raw source from `raw/sources/` (and `raw/assets/` references if needed).
-2. Extract key claims, facts, entities, concepts, uncertainty.
-3. Create/update one source summary in `wiki/sources/`.
-4. Update relevant `wiki/entities/*` and `wiki/concepts/*` pages.
-5. Update `wiki/overview.md` synthesis if this source materially changes understanding.
-6. Update `index.md`.
-7. Append ingest entry to `log.md`.
-8. Report what changed, what is uncertain, and suggested next source/questions.
-9. If this ingest created or modified any page in `wiki/concepts/`, run Workflow D
-   (Sync Script Coverage) before reporting.
-
-### Workflow B: Answer Query
-
-1. Read `index.md` first.
-2. Select relevant pages and synthesize answer with page citations.
-3. If answer has durable value, save to `wiki/queries/YYYY-MM-DD-<slug>.md`.
-4. If answer reveals new synthesis, update affected concept/entity pages.
-5. Update `index.md` and append query entry in `log.md` when files changed.
-
-### Workflow C: Lint Wiki
-
-Run periodic health checks for:
-
-- contradiction detection across pages
-- stale claims superseded by newer sources
-- orphan pages / weak linking
-- high-mention concepts lacking dedicated pages
-- missing evidence links
-- missing or folder-mismatched page-type tags (per Tagging Rules)
-
-Write report to `wiki/lint-reports/YYYY-MM-DD-lint.md`, then update `index.md` and `log.md`.
-
-### Workflow D: Sync Script Coverage
-
-Maintains `wiki/script-coverage.md` — one row per page in `wiki/concepts/`, judged
-against the script named in that file's `**Script:**` metadata field.
-
-1. Read `wiki/script-coverage.md` and note every row where `Pinned` is `yes`.
-2. Read the script and every page in `wiki/concepts/`.
-3. For each concept, decide `Status` and `Station`:
-   - `covered` — the script delivers the idea, whether or not it uses the page's name.
-   - `partial` — the script gestures at it but never lands it.
-   - `absent` — the script never reaches it.
-   - `Station` is one of the seven — `Chat box`, `ReAct`, `Tools`, `Memory`,
-     `Skills`, `Process`, `OS` — spelled exactly as written here, or a
-     comma-separated list of them, or the lowercase word `all`, or `—` when absent.
-4. **Never modify a row whose `Pinned` is `yes`** — not its status, not its station.
-   A pinned row is the user's judgment and outranks yours.
-5. Add rows for concept pages with no row. Remove rows whose concept page no longer exists.
-   A pinned row whose concept page no longer exists is still removed — step 4
-   protects the user's judgement about a live concept, not a dangling row.
-6. Update `**Last synced:**` to today.
-7. Update `index.md` and append a `sync` entry to `log.md`.
-
-Run this workflow:
-
-- at the end of any ingest that creates or modifies a page in `wiki/concepts/`
-- whenever the script file itself changes
-- on the explicit `sync script coverage` intent
-
-The coverage baseline is always the raw script. Ingesting the script into
-`wiki/sources/` does not change the baseline.
-
-## Citation & Evidence Policy
-
-- Prefer citing wiki source summaries (`wiki/sources/*.md`) rather than raw files in normal answers.
-- If citing raw material directly, also reflect it into a source summary page.
-- Mark uncertain or contested claims with `Status: tentative` in relevant section.
-
-## Update Policy
-
-- Never perform silent large rewrites.
-- Preserve meaningful prior interpretations; move outdated material under a "Superseded" note when needed.
-- Keep sections ordered consistently for predictable diffs.
-
-## Operational Commands (Natural Language)
-
-Supported intents:
-
-- "ingest path-or-title"
-- "query: question"
-- "lint wiki"
-- "show recent changes"
-- "suggest next sources"
-- "sync script coverage"
-
-Always execute intents according to workflows above.
-
-## Session Start Checklist
-
-At start of every session:
-
-1. Read `index.md` and the latest section of `log.md`.
-2. Identify last completed operation and current open questions.
-3. Continue from prior state without resetting conventions.
+# LLM Wiki Schema (Operational Contract)
+
+You are the dedicated maintainer of this vault as a persistent, compounding wiki.
+
+## Mission
+
+Maintain a high-signal personal knowledge base where:
+
+- `raw/` contains immutable source materials.
+- `wiki/` contains LLM-authored, continuously maintained markdown pages.
+- `index.md` is the content catalog.
+- `log.md` is an append-only chronological operation log.
+
+The user curates sources and asks questions. You perform all wiki maintenance.
+
+## Non-Negotiable Rules
+
+1. Never modify files inside `raw/`.
+2. Always update `index.md` and `log.md` after any ingest/query/lint operation that changes the wiki.
+3. Prefer editing existing pages over creating duplicates.
+4. Use Obsidian-style wiki links (`[[page-name]]`) for internal references.
+5. Record contradictions and superseded claims explicitly (do not silently overwrite history).
+6. Keep claims attributable: include source links to `wiki/sources/*` pages.
+7. Do not leave orphan pages intentionally; add at least one inbound and one outbound link when possible.
+8. Keep writing concise, structured, and diff-friendly.
+9. Every wiki page carries exactly one page-type hashtag (see Tagging Rules) matching its folder.
+
+## Folder Convention
+
+```text
+raw/
+  sources/            # immutable source markdown/text/pdf exports
+  assets/             # immutable local images/files referenced by raw sources
+
+wiki/
+  overview.md         # top-level synthesis and navigation
+  script-coverage.md  # machine-maintained: concept coverage vs the webinar script
+  sources/            # one summary page per ingested raw source
+  entities/           # people, orgs, projects, places, tools
+  concepts/           # themes, ideas, methods, frameworks
+  timelines/          # optional chronological reconstructions
+  comparisons/        # side-by-side analyses
+  queries/            # durable outputs created from Q&A sessions
+  lint-reports/       # periodic health-check reports
+```
+
+## File Naming Rules
+
+- Use kebab-case for file names.
+- Prefix source summary pages with date: `YYYY-MM-DD-title.md`.
+- Prefer stable canonical pages:
+  - `wiki/entities/<name>.md`
+  - `wiki/concepts/<concept>.md`
+- If a page name collides, merge instead of creating suffixes unless truly distinct.
+
+## Tagging Rules
+
+Every page in `wiki/` starts with its page-type hashtag on its own line, directly after the `# <title>` heading (Obsidian inline tag format):
+
+```text
+# <Title>
+
+#<type-tag>
+
+## <first section>
+```
+
+Folder → required tag:
+
+| Folder / file          | Tag            |
+| ---------------------- | -------------- |
+| `wiki/overview.md`     | `#overview`    |
+| `wiki/script-coverage.md` | `#coverage`    |
+| `wiki/sources/*`       | `#source`      |
+| `wiki/entities/*`      | `#entity`      |
+| `wiki/concepts/*`      | `#concept`     |
+| `wiki/timelines/*`     | `#timeline`    |
+| `wiki/comparisons/*`   | `#comparison`  |
+| `wiki/queries/*`       | `#query`       |
+| `wiki/lint-reports/*`  | `#lint-report` |
+
+- Exactly one page-type tag per page; it must match the page's folder.
+- Additional topical tags are allowed on the same line after the type tag (e.g. `#concept #ai-tooling`), but the type tag comes first.
+- When creating any new page, add the tag immediately; when moving a page between folders, update the tag in the same operation.
+
+## Required Page Templates
+
+### 1) Source Summary (`wiki/sources/*.md`)
+
+Must include sections:
+
+- `# <title>`
+- `#source` tag line (per Tagging Rules)
+- `## Source Metadata` (date, raw path, source type, ingestion date)
+- `## Core Claims`
+- `## Key Evidence / Details`
+- `## Connections` (links to entities/concepts/comparisons)
+- `## Open Questions`
+- `## Change Impact on Wiki` (what pages were updated and why)
+
+### 2) Entity / Concept Page
+
+Must include:
+
+- `# <name>`
+- `#entity` or `#concept` tag line (per Tagging Rules)
+- `## Summary`
+- `## Current Understanding`
+- `## Evidence` (bullets with links to `wiki/sources/*`)
+- `## Related Pages`
+- `## Contradictions / Uncertainty`
+- `## Next Questions`
+
+### 3) Query Output (`wiki/queries/*.md`)
+
+Must include:
+
+- `#query` tag line (per Tagging Rules)
+- Question asked
+- Answer
+- Evidence trail (links)
+- Follow-up questions
+- Whether this output changed existing pages
+
+### 4) Coverage File (`wiki/script-coverage.md`)
+
+Machine-read by the dashboard plugin, which parses the table positionally. The
+shape below is required exactly — a deviation renders as parse errors, not as a
+best-effort read.
+
+- `# Script coverage`
+- `#coverage` tag line
+- `## Metadata`, containing these two lines verbatim in this form:
+  - `- **Script:** ` followed by the raw script path in backticks
+  - `- **Last synced:** ` followed by a `YYYY-MM-DD` date
+- `## Coverage`, containing one markdown table with exactly these four columns in
+  this order: `Concept`, `Status`, `Station`, `Pinned`
+  - **Concept** — an Obsidian wikilink to a page in `wiki/concepts/`
+  - **Status** — exactly one of `covered`, `partial`, `absent`, lowercase
+  - **Station** — one of the seven station names spelled exactly as listed in
+    Workflow D, a comma-separated list of them, the lowercase word `all`, or an
+    em dash `—` when the status is `absent`
+  - **Pinned** — the lowercase word `yes`, or empty
+- `## Notes` and `## Related Pages` are free prose and are not parsed.
+
+The parser reads every markdown table in the file, so do not add a second table.
+
+## Standard Workflows
+
+### Workflow A: Ingest One Source
+
+When user says "ingest <source>":
+
+1. Read the raw source from `raw/sources/` (and `raw/assets/` references if needed).
+2. Extract key claims, facts, entities, concepts, uncertainty.
+3. Create/update one source summary in `wiki/sources/`.
+4. Update relevant `wiki/entities/*` and `wiki/concepts/*` pages.
+5. Update `wiki/overview.md` synthesis if this source materially changes understanding.
+6. Update `index.md`.
+7. Append ingest entry to `log.md`.
+8. Report what changed, what is uncertain, and suggested next source/questions.
+9. If this ingest created or modified any page in `wiki/concepts/`, run Workflow D
+   (Sync Script Coverage) before reporting.
+
+### Workflow B: Answer Query
+
+1. Read `index.md` first.
+2. Select relevant pages and synthesize answer with page citations.
+3. If answer has durable value, save to `wiki/queries/YYYY-MM-DD-<slug>.md`.
+4. If answer reveals new synthesis, update affected concept/entity pages.
+5. Update `index.md` and append query entry in `log.md` when files changed.
+
+### Workflow C: Lint Wiki
+
+Run periodic health checks for:
+
+- contradiction detection across pages
+- stale claims superseded by newer sources
+- orphan pages / weak linking
+- high-mention concepts lacking dedicated pages
+- missing evidence links
+- missing or folder-mismatched page-type tags (per Tagging Rules)
+
+Write report to `wiki/lint-reports/YYYY-MM-DD-lint.md`, then update `index.md` and `log.md`.
+
+### Workflow D: Sync Script Coverage
+
+Maintains `wiki/script-coverage.md` — one row per page in `wiki/concepts/`, judged
+against the script named in that file's `**Script:**` metadata field.
+
+1. Read `wiki/script-coverage.md` and note every row where `Pinned` is `yes`.
+2. Read the script and every page in `wiki/concepts/`.
+3. For each concept, decide `Status` and `Station`:
+   - `covered` — the script delivers the idea, whether or not it uses the page's name.
+   - `partial` — the script gestures at it but never lands it.
+   - `absent` — the script never reaches it.
+   - `Station` is one of the seven — `Chat box`, `ReAct`, `Tools`, `Memory`,
+     `Skills`, `Process`, `OS` — spelled exactly as written here, or a
+     comma-separated list of them, or the lowercase word `all`, or `—` when absent.
+4. **Never modify a row whose `Pinned` is `yes`** — not its status, not its station.
+   A pinned row is the user's judgment and outranks yours.
+5. Add rows for concept pages with no row. Remove rows whose concept page no longer exists.
+   A pinned row whose concept page no longer exists is still removed — step 4
+   protects the user's judgement about a live concept, not a dangling row.
+6. Update `**Last synced:**` to today.
+7. Update `index.md` and append a `sync` entry to `log.md`.
+
+Run this workflow:
+
+- at the end of any ingest that creates or modifies a page in `wiki/concepts/`
+- whenever the script file itself changes
+- on the explicit `sync script coverage` intent
+
+The coverage baseline is always the raw script. Ingesting the script into
+`wiki/sources/` does not change the baseline.
+
+## Citation & Evidence Policy
+
+- Prefer citing wiki source summaries (`wiki/sources/*.md`) rather than raw files in normal answers.
+- If citing raw material directly, also reflect it into a source summary page.
+- Mark uncertain or contested claims with `Status: tentative` in relevant section.
+
+## Update Policy
+
+- Never perform silent large rewrites.
+- Preserve meaningful prior interpretations; move outdated material under a "Superseded" note when needed.
+- Keep sections ordered consistently for predictable diffs.
+
+## Operational Commands (Natural Language)
+
+Supported intents:
+
+- "ingest path-or-title"
+- "query: question"
+- "lint wiki"
+- "show recent changes"
+- "suggest next sources"
+- "sync script coverage"
+
+Always execute intents according to workflows above.
+
+## Session Start Checklist
+
+At start of every session:
+
+1. Read `index.md` and the latest section of `log.md`.
+2. Identify last completed operation and current open questions.
+3. Continue from prior state without resetting conventions.
diff --git a/docs/superpowers/plans/2026-07-28-webinar-dashboard.md b/docs/superpowers/plans/2026-07-28-webinar-dashboard.md
index 4384297..9391cdc 100644
--- a/docs/superpowers/plans/2026-07-28-webinar-dashboard.md
+++ b/docs/superpowers/plans/2026-07-28-webinar-dashboard.md
@@ -1,1745 +1,1745 @@
-# Webinar Vault Dashboard Implementation Plan
-
-> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
-
-**Goal:** Build an Obsidian plugin that renders a two-pane dashboard showing the vault's source pipeline (with one-click headless ingest) and concept coverage against the webinar script.
-
-**Architecture:** A single unbundled `main.js` CommonJS plugin. Pure logic (path extraction, pipeline derivation, table parsing, filename validation) lives at the top of the file behind a guarded `require("obsidian")`, so it can be unit-tested with plain `node --test` and no Obsidian runtime. Obsidian glue (code-block processor, DOM rendering, subprocess spawn) sits below it. The plugin reads the vault and never writes to it.
-
-**Tech Stack:** Plain CommonJS (no TypeScript, no bundler, no build step), Node's built-in `node:test` runner, Node `child_process`, Obsidian Plugin API, tesanti design tokens.
-
-## Global Constraints
-
-- **No build step.** `main.js` is loaded by Obsidian verbatim. No TypeScript, no esbuild, no `npm install`.
-- **No dependencies.** Tests use Node's built-in `node:test` and `node:assert/strict` only.
-- **`isDesktopOnly: true`** is mandatory in `manifest.json` — the plugin uses `child_process`, which Obsidian only provides on desktop.
-- **`minAppVersion`: `"1.5.0"`.**
-- **Relative `require()` between plugin files is unsupported.** All code lives in one `main.js`.
-- **The plugin never writes to the vault.** It reads files and spawns one subprocess. It must never read `index.md`.
-- **Design tokens are copied verbatim** from `tesanti Design System.zip → colors_and_type.css`. Black / white / red only; radii at most 6px; 1px hairlines instead of shadows.
-- **No emoji anywhere** — in code, UI, comments, or commit messages. Status icons are inline Lucide stroke paths.
-- **Sentence case** for all UI copy. No Title Case.
-- **The seven stations**, in order: `Chat box`, `ReAct`, `Tools`, `Memory`, `Skills`, `Process`, `OS`.
-- **Valid statuses**, exactly: `covered`, `partial`, `absent`.
-
----
-
-## File Structure
-
-| File | Responsibility |
-|---|---|
-| `.obsidian/plugins/webinar-dash/manifest.json` | Plugin identity and desktop-only flag |
-| `.obsidian/plugins/webinar-dash/main.js` | Everything: pure helpers, renderers, plugin class |
-| `.obsidian/plugins/webinar-dash/styles.css` | tesanti tokens and dashboard styling, scoped to `.webinar-dash` |
-| `.obsidian/plugins/webinar-dash/test/pipeline.test.js` | Tests for `extractRawPath`, `derivePipeline` |
-| `.obsidian/plugins/webinar-dash/test/coverage.test.js` | Tests for `parseCoverageTable`, `groupByStation` |
-| `.obsidian/plugins/webinar-dash/test/safety.test.js` | Tests for `isSafeFilename` |
-| `dashboard.md` | Vault root. Holds one `webinar-dash` config block |
-| `wiki/script-coverage.md` | The coverage table |
-| `CLAUDE.md` | Modified: folder convention, tagging rules, Workflow D, sync triggers |
-
-Tests live inside the plugin folder. Obsidian loads only `main.js` from a plugin directory, so `test/` is inert at runtime.
-
----
-
-### Task 1: Repository and plugin scaffold
-
-**Files:**
-- Create: `.gitignore`
-- Create: `.obsidian/plugins/webinar-dash/manifest.json`
-- Create: `.obsidian/plugins/webinar-dash/main.js`
-- Create: `dashboard.md`
-
-**Interfaces:**
-- Consumes: nothing
-- Produces: a loadable plugin registering the `webinar-dash` code-block language; `module.exports.__test__` as the export surface every later task extends
-
-- [ ] **Step 1: Initialize the repository**
-
-The vault is not currently a git repository. The plugin will spawn agents that write to `wiki/`, `index.md`, and `log.md` unsupervised, so an undo path is required before that capability exists.
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git init
-```
-
-- [ ] **Step 2: Create `.gitignore`**
-
-Obsidian rewrites `workspace.json` constantly; it is local UI state, not vault content.
-
-```gitignore
-.obsidian/workspace.json
-.obsidian/workspace-mobile.json
-.obsidian/cache
-```
-
-- [ ] **Step 3: Commit the vault as it stands**
-
-```bash
-git add -A
-git commit -m "chore: initial commit of vault before dashboard work"
-```
-
-- [ ] **Step 4: Create the manifest**
-
-`isDesktopOnly` must be `true` — Obsidian only exposes `child_process` on desktop, and the ingest button depends on it.
-
-```json
-{
-  "id": "webinar-dash",
-  "name": "Webinar dashboard",
-  "version": "0.1.0",
-  "minAppVersion": "1.5.0",
-  "description": "Source pipeline and script coverage for the webinar vault.",
-  "author": "meels",
-  "isDesktopOnly": true
-}
-```
-
-- [ ] **Step 5: Create `main.js` with the Obsidian guard and a stub renderer**
-
-The `try/catch` around `require("obsidian")` is what lets `node --test` load this file. Under Node the require throws, `OB` stays null, and `PluginBase` becomes an empty class so `class ... extends PluginBase` still evaluates. Every later task adds pure functions above the plugin class and registers them in `__test__`.
-
-```js
-"use strict";
-
-// Obsidian injects its own module resolver. Under plain `node --test` it is
-// absent, so guard the require and fall back to an empty base class. This is
-// what keeps the pure helpers below unit-testable without an Obsidian runtime.
-let OB = null;
-try {
-  OB = require("obsidian");
-} catch (_) {
-  OB = null;
-}
-const PluginBase = OB ? OB.Plugin : class {};
-
-const STATIONS = ["Chat box", "ReAct", "Tools", "Memory", "Skills", "Process", "OS"];
-
-const DEFAULTS = {
-  script: "raw/sources/Webinar script.md",
-  coverage: "wiki/script-coverage.md",
-  rawDir: "raw/sources",
-  wikiSourceDir: "wiki/sources",
-  conceptDir: "wiki/concepts",
-};
-
-function parseConfig(source) {
-  const cfg = Object.assign({}, DEFAULTS);
-  for (const line of String(source).split(/\r?\n/)) {
-    const m = line.match(/^\s*([A-Za-z][A-Za-z0-9_]*)\s*:\s*(.+?)\s*$/);
-    if (m && Object.prototype.hasOwnProperty.call(DEFAULTS, m[1])) {
-      cfg[m[1]] = m[2];
-    }
-  }
-  return cfg;
-}
-
-class WebinarDashPlugin extends PluginBase {
-  async onload() {
-    this.registerMarkdownCodeBlockProcessor("webinar-dash", (source, el, ctx) => {
-      const cfg = parseConfig(source);
-      const root = el.createDiv({ cls: "webinar-dash" });
-      root.createDiv({ cls: "wd-eyebrow", text: "Webinar dashboard" });
-      root.createEl("p", { text: `Reading coverage from ${cfg.coverage}` });
-    });
-  }
-}
-
-module.exports = WebinarDashPlugin;
-module.exports.default = WebinarDashPlugin;
-module.exports.__test__ = { parseConfig, STATIONS, DEFAULTS };
-```
-
-- [ ] **Step 6: Create `dashboard.md`**
-
-````markdown
-# Webinar dashboard
-
-```webinar-dash
-script: raw/sources/Webinar script.md
-coverage: wiki/script-coverage.md
-```
-````
-
-- [ ] **Step 7: Enable the plugin and verify it renders**
-
-In Obsidian: Settings → Community plugins → turn off Restricted mode if on → Installed plugins → Reload → enable "Webinar dashboard". Open `dashboard.md` in reading view.
-
-Expected: the text `Webinar dashboard` followed by `Reading coverage from wiki/script-coverage.md`.
-
-If nothing renders, run "Reload app without saving" (`Ctrl+R`) — Obsidian caches plugin code between edits.
-
-- [ ] **Step 8: Commit**
-
-```bash
-git add .gitignore .obsidian/plugins/webinar-dash dashboard.md
-git commit -m "feat: scaffold webinar-dash plugin with config block"
-```
-
----
-
-### Task 2: Source pipeline derivation
-
-**Files:**
-- Modify: `.obsidian/plugins/webinar-dash/main.js`
-- Test: `.obsidian/plugins/webinar-dash/test/pipeline.test.js`
-
-**Interfaces:**
-- Consumes: `module.exports.__test__` from Task 1
-- Produces:
-  - `extractRawPath(text: string) => string | null`
-  - `derivePipeline({ rawFiles, sourcePages }) => { processed, unprocessed, orphaned }`
-  - `rawFiles` items: `{ path: string, name: string, size: number }`
-  - `sourcePages` items: `{ path: string, name: string, rawPath: string | null }`
-  - `processed` items: `rawFiles` item plus `{ page }`
-
-- [ ] **Step 1: Write the failing tests**
-
-Create `.obsidian/plugins/webinar-dash/test/pipeline.test.js`. The fixtures are the real strings from this vault.
-
-```js
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { extractRawPath, derivePipeline } = require("../main.js").__test__;
-
-test("extractRawPath pulls the backticked path", () => {
-  const page = [
-    "# You're reading way too much code",
-    "",
-    "#source",
-    "",
-    "## Source Metadata",
-    "",
-    "- **Date:** YouTube video, 24:11",
-    "- **Raw path:** `raw/sources/You're reading way too much code.md`",
-    "- **Source type:** video essay",
-  ].join("\n");
-  assert.equal(extractRawPath(page), "raw/sources/You're reading way too much code.md");
-});
-
-test("extractRawPath handles cyrillic and em dashes", () => {
-  const page = "- **Raw path:** `raw/sources/Скиллы на базе git — новая память AI-агентов.md`";
-  assert.equal(extractRawPath(page), "raw/sources/Скиллы на базе git — новая память AI-агентов.md");
-});
-
-test("extractRawPath returns null when the line is absent", () => {
-  assert.equal(extractRawPath("# A page\n\n#source\n\nNo metadata here."), null);
-});
-
-test("derivePipeline splits claimed from unclaimed raw files", () => {
-  const rawFiles = [
-    { path: "raw/sources/Nina interview.md", name: "Nina interview.md", size: 6614 },
-    { path: "raw/sources/Webinar script.md", name: "Webinar script.md", size: 15841 },
-  ];
-  const sourcePages = [
-    {
-      path: "wiki/sources/2026-07-14-nina-interview.md",
-      name: "2026-07-14-nina-interview.md",
-      rawPath: "raw/sources/Nina interview.md",
-    },
-  ];
-  const out = derivePipeline({ rawFiles, sourcePages });
-  assert.equal(out.processed.length, 1);
-  assert.equal(out.processed[0].name, "Nina interview.md");
-  assert.equal(out.processed[0].page.name, "2026-07-14-nina-interview.md");
-  assert.equal(out.unprocessed.length, 1);
-  assert.equal(out.unprocessed[0].name, "Webinar script.md");
-  assert.equal(out.orphaned.length, 0);
-});
-
-test("derivePipeline reports source pages whose raw file is gone", () => {
-  const out = derivePipeline({
-    rawFiles: [],
-    sourcePages: [
-      { path: "wiki/sources/x.md", name: "x.md", rawPath: "raw/sources/deleted.md" },
-    ],
-  });
-  assert.equal(out.orphaned.length, 1);
-  assert.equal(out.orphaned[0].name, "x.md");
-});
-
-test("derivePipeline treats a page with no raw path as orphaned", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
-    sourcePages: [{ path: "wiki/sources/y.md", name: "y.md", rawPath: null }],
-  });
-  assert.equal(out.orphaned.length, 1);
-  assert.equal(out.unprocessed.length, 1);
-});
-
-test("derivePipeline routes a duplicate raw-path claim to orphaned", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
-    sourcePages: [
-      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/a.md" },
-      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/a.md" },
-    ],
-  });
-  assert.equal(out.processed.length, 1);
-  assert.equal(out.processed[0].page.name, "first.md");
-  assert.equal(out.unprocessed.length, 0);
-  assert.deepEqual(out.orphaned.map((p) => p.name), ["second.md"]);
-});
-
-test("derivePipeline lists a page once when it is both a duplicate claim and missing its raw file", () => {
-  const out = derivePipeline({
-    rawFiles: [{ path: "raw/sources/other.md", name: "other.md", size: 10 }],
-    sourcePages: [
-      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/gone.md" },
-      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/gone.md" },
-    ],
-  });
-  assert.deepEqual(out.orphaned.map((p) => p.name), ["first.md", "second.md"]);
-});
-
-test("derivePipeline sorts unprocessed by name and processed newest first", () => {
-  const out = derivePipeline({
-    rawFiles: [
-      { path: "raw/sources/b.md", name: "b.md", size: 1 },
-      { path: "raw/sources/a.md", name: "a.md", size: 1 },
-      { path: "raw/sources/c.md", name: "c.md", size: 1 },
-    ],
-    sourcePages: [
-      { path: "wiki/sources/2026-07-14-x.md", name: "2026-07-14-x.md", rawPath: "raw/sources/c.md" },
-    ],
-  });
-  assert.deepEqual(out.unprocessed.map((f) => f.name), ["a.md", "b.md"]);
-  assert.equal(out.processed.length, 1);
-});
-```
-
-- [ ] **Step 2: Run the tests to verify they fail**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test test/pipeline.test.js
-```
-
-Expected: FAIL — `extractRawPath is not a function` (it is not yet exported).
-
-- [ ] **Step 3: Implement the helpers**
-
-In `main.js`, insert directly below the `DEFAULTS` constant:
-
-```js
-const RAW_PATH_RE = /^\s*-\s*\*\*Raw path:\*\*\s*`([^`]+)`/m;
-
-function extractRawPath(text) {
-  const m = String(text).match(RAW_PATH_RE);
-  return m ? m[1].trim() : null;
-}
-
-function derivePipeline({ rawFiles, sourcePages }) {
-  // First claim on a raw path wins. A later page claiming the same file is a
-  // duplicate claim — real catalog drift — and joins `orphaned` rather than
-  // being silently dropped. `orphaned` therefore means "source page not paired
-  // with a raw file", whatever the reason.
-  const claimed = new Map();
-  const duplicates = [];
-  for (const page of sourcePages) {
-    if (!page.rawPath) continue;
-    if (claimed.has(page.rawPath)) duplicates.push(page);
-    else claimed.set(page.rawPath, page);
-  }
-
-  const processed = [];
-  const unprocessed = [];
-  for (const file of rawFiles) {
-    const page = claimed.get(file.path);
-    if (page) processed.push(Object.assign({}, file, { page }));
-    else unprocessed.push(file);
-  }
-
-  // One pass over sourcePages, so a page appears in `orphaned` at most once no
-  // matter how many of the three reasons apply to it. Concatenating a separate
-  // duplicates array here would double-count a losing claimant whose shared raw
-  // path is also missing from disk.
-  const rawPaths = new Set(rawFiles.map((f) => f.path));
-  const duplicateSet = new Set(duplicates);
-  const orphaned = sourcePages.filter(
-    (p) => duplicateSet.has(p) || !p.rawPath || !rawPaths.has(p.rawPath)
-  );
-
-  unprocessed.sort((a, b) => a.name.localeCompare(b.name));
-  processed.sort((a, b) => b.page.name.localeCompare(a.page.name));
-
-  return { processed, unprocessed, orphaned };
-}
-```
-
-- [ ] **Step 4: Export them**
-
-Replace the `__test__` line at the bottom of `main.js`:
-
-```js
-module.exports.__test__ = { parseConfig, extractRawPath, derivePipeline, STATIONS, DEFAULTS };
-```
-
-- [ ] **Step 5: Run the tests to verify they pass**
-
-```bash
-node --test test/pipeline.test.js
-```
-
-Expected: PASS, 9 tests.
-
-- [ ] **Step 6: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add .obsidian/plugins/webinar-dash
-git commit -m "feat: derive source pipeline from raw path claims"
-```
-
----
-
-### Task 3: Left pane — source pipeline rendering
-
-**Files:**
-- Modify: `.obsidian/plugins/webinar-dash/main.js`
-- Create: `.obsidian/plugins/webinar-dash/styles.css`
-
-**Interfaces:**
-- Consumes: `derivePipeline`, `extractRawPath` from Task 2; `parseConfig` from Task 1
-- Produces:
-  - `async readPipeline(app, cfg) => { processed, unprocessed, orphaned }` as a method on the plugin class
-  - `renderLeftPane(container, pipeline, onIngest)` where `onIngest` is `(file, rowEl) => void`; Task 6 supplies the real handler, this task passes a no-op
-  - CSS class root `.webinar-dash` and the shared primitives `.wd-eyebrow`, `.wd-row`, `.wd-btn`, `.wd-mono`
-
-- [ ] **Step 1: Create `styles.css` with tesanti tokens**
-
-Values are copied verbatim from `colors_and_type.css`. Scoped to `.webinar-dash` so nothing leaks into the rest of Obsidian. Obsidian loads `styles.css` from the plugin folder automatically.
-
-```css
-.webinar-dash {
-  --wd-ink-000: #ffffff; --wd-ink-050: #f7f7f7; --wd-ink-100: #ececec;
-  --wd-ink-200: #d9d9d9; --wd-ink-400: #8a8a8a; --wd-ink-500: #5e5e5e;
-  --wd-ink-700: #262626; --wd-ink-900: #0a0a0a; --wd-ink-999: #000000;
-  --wd-red-500: #e1261c; --wd-red-600: #c31c14;
-
-  --wd-bg: var(--wd-ink-000);
-  --wd-bg-subtle: var(--wd-ink-050);
-  --wd-fg: var(--wd-ink-999);
-  --wd-fg-2: var(--wd-ink-700);
-  --wd-fg-3: var(--wd-ink-500);
-  --wd-fg-4: var(--wd-ink-400);
-  --wd-border: var(--wd-ink-200);
-  --wd-border-strong: var(--wd-ink-999);
-  --wd-border-subtle: var(--wd-ink-100);
-  --wd-accent: var(--wd-red-500);
-  --wd-accent-press: var(--wd-red-600);
-  --wd-accent-on: #ffffff;
-  --wd-ok: #0a8a3f;
-  --wd-warn: #c68a00;
-  --wd-danger: var(--wd-red-500);
-
-  --wd-mono: "JetBrains Mono", ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
-  --wd-sans: "Inter", ui-sans-serif, system-ui, -apple-system, "Helvetica Neue", Arial, sans-serif;
-
-  font-family: var(--wd-sans);
-  color: var(--wd-fg);
-  display: grid;
-  grid-template-columns: minmax(0, 38fr) minmax(0, 62fr);
-  gap: 24px;
-}
-
-.theme-dark .webinar-dash {
-  --wd-bg: var(--wd-ink-900);
-  --wd-bg-subtle: #161616;
-  --wd-fg: var(--wd-ink-000);
-  --wd-fg-2: var(--wd-ink-200);
-  --wd-fg-3: var(--wd-ink-400);
-  --wd-fg-4: var(--wd-ink-500);
-  --wd-border: var(--wd-ink-700);
-  --wd-border-strong: #b8b8b8;
-  --wd-border-subtle: #161616;
-  /* Accent tracks the installed theme at .obsidian/themes/tesanti/theme.css,
-     whose dark section reads "Black canvas, same signal red": --accent-h/-s/-l
-     are declared once at :root and never redeclared under .theme-dark, and only
-     the hover state lifts (#c31c14 light, #ff4d43 dark). Match that exactly. */
-  --wd-accent: var(--wd-red-500);
-  --wd-accent-press: #ff4d43;
-  --wd-accent-on: #ffffff;
-  /* Status tokens are separate from the brand accent by design-system rule, and
-     here they carry small text in a dense table. #e1261c on near-black is about
-     3.8:1, under AA for small text, so these lift where the accent does not. */
-  --wd-ok: #2fbf6a;
-  --wd-warn: #e0a516;
-  --wd-danger: #ff5c50;
-}
-
-@media (max-width: 820px) {
-  .webinar-dash { grid-template-columns: minmax(0, 1fr); }
-}
-
-.webinar-dash > * { min-width: 0; }
-
-.wd-pane { display: flex; flex-direction: column; gap: 20px; }
-
-.wd-eyebrow {
-  font-family: var(--wd-mono); font-size: 11px; line-height: 1;
-  letter-spacing: 0.12em; text-transform: uppercase;
-  color: var(--wd-fg-3); font-weight: 500;
-  display: flex; align-items: center; gap: 8px;
-}
-.wd-eyebrow::before {
-  content: ""; width: 5px; height: 5px; flex: none; background: var(--wd-accent);
-}
-
-.wd-block { display: flex; flex-direction: column; gap: 12px; }
-
-.wd-row {
-  display: flex; align-items: center; gap: 12px;
-  border: 1px solid var(--wd-border); border-radius: 4px;
-  padding: 12px 12px 12px 16px; background: var(--wd-bg);
-}
-.wd-row + .wd-row { margin-top: -1px; }
-.wd-row-main { flex: 1; min-width: 0; display: flex; flex-direction: column; gap: 2px; }
-.wd-row-name {
-  font-size: 14px; font-weight: 600;
-  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
-}
-.wd-mono {
-  font-family: var(--wd-mono); font-size: 11px; color: var(--wd-fg-3);
-  font-variant-numeric: tabular-nums;
-}
-
-.wd-btn {
-  display: inline-flex; align-items: center; gap: 6px; flex: none;
-  font-family: var(--wd-sans); font-size: 13px; font-weight: 600;
-  padding: 6px 12px; border-radius: 6px; cursor: pointer;
-  border: 1px solid var(--wd-accent); background: var(--wd-accent);
-  color: var(--wd-accent-on);
-  transition: background 120ms cubic-bezier(0.2, 0, 0, 1);
-}
-.wd-btn:hover { background: var(--wd-accent-press); border-color: var(--wd-accent-press); }
-.wd-btn:focus-visible { outline: 2px solid var(--wd-accent); outline-offset: 2px; }
-.wd-btn[disabled] {
-  opacity: 0.45; cursor: not-allowed;
-  background: transparent; color: var(--wd-fg-3); border-color: var(--wd-border);
-}
-.wd-btn svg { width: 14px; height: 14px; flex: none; }
-
-.wd-done { display: flex; flex-direction: column; }
-.wd-done-row {
-  display: flex; align-items: baseline; gap: 12px; padding: 5px 0;
-  border-bottom: 1px solid var(--wd-border-subtle); font-size: 13px;
-}
-.wd-done-row:last-child { border-bottom: 0; }
-.wd-done-name {
-  color: var(--wd-fg-2);
-  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
-}
-
-.wd-note { font-size: 12px; color: var(--wd-fg-3); }
-.wd-error {
-  font-size: 13px; color: var(--wd-danger);
-  border: 1px solid var(--wd-danger); border-radius: 4px; padding: 12px;
-}
-
-@media (prefers-reduced-motion: reduce) {
-  .webinar-dash * { transition-duration: 0.01ms !important; }
-}
-```
-
-- [ ] **Step 2: Add the Lucide icon helper and the left-pane renderer**
-
-Insert into `main.js` above the plugin class. Icons are inline stroke paths — the Lucide CDN is unavailable offline and the design system forbids emoji.
-
-```js
-// Source types CLAUDE.md documents for raw/sources. Anything else in that
-// folder is not a source and stays out of the queue.
-const RAW_EXTENSIONS = new Set(["md", "txt", "pdf"]);
-
-const ICONS = {
-  terminal: "M4 17l6-6-6-6M12 19h8",
-  check: "M20 6 9 17l-5-5",
-  minus: "M5 12h14",
-  x: "M18 6 6 18M6 6l12 12",
-  refresh: "M3 12a9 9 0 0 1 9-9 9 9 0 0 1 6.7 3L21 8M21 3v5h-5 M21 12a9 9 0 0 1-9 9 9 9 0 0 1-6.7-3L3 16M3 21v-5h5",
-};
-
-// Built with createElementNS rather than Obsidian's createSvg helper, whose
-// availability varies by version. This works on any Obsidian build.
-const SVG_NS = "http://www.w3.org/2000/svg";
-
-function addIcon(parent, name) {
-  const svg = document.createElementNS(SVG_NS, "svg");
-  svg.setAttribute("viewBox", "0 0 24 24");
-  svg.setAttribute("fill", "none");
-  svg.setAttribute("aria-hidden", "true");
-  const path = document.createElementNS(SVG_NS, "path");
-  path.setAttribute("d", ICONS[name]);
-  path.setAttribute("stroke", "currentColor");
-  path.setAttribute("stroke-width", "1.75");
-  path.setAttribute("stroke-linecap", "round");
-  path.setAttribute("stroke-linejoin", "round");
-  svg.appendChild(path);
-  parent.appendChild(svg);
-  return svg;
-}
-
-function formatBytes(n) {
-  return `${n.toLocaleString("en-US")} B`;
-}
-
-function renderLeftPane(container, pipeline, onIngest) {
-  const pane = container.createDiv({ cls: "wd-pane" });
-
-  const queue = pane.createDiv({ cls: "wd-block" });
-  queue.createDiv({
-    cls: "wd-eyebrow",
-    text: `Queue — ${pipeline.unprocessed.length} unprocessed`,
-  });
-  if (pipeline.unprocessed.length === 0) {
-    queue.createDiv({ cls: "wd-note", text: "Every raw source has a summary page." });
-  }
-  for (const file of pipeline.unprocessed) {
-    const row = queue.createDiv({ cls: "wd-row" });
-    const main = row.createDiv({ cls: "wd-row-main" });
-    main.createDiv({ cls: "wd-row-name", text: file.name });
-    main.createDiv({ cls: "wd-mono", text: formatBytes(file.size) });
-    const btn = row.createEl("button", { cls: "wd-btn" });
-    addIcon(btn, "terminal");
-    btn.createSpan({ text: "Ingest" });
-    btn.addEventListener("click", () => onIngest(file, row));
-  }
-
-  const done = pane.createDiv({ cls: "wd-block" });
-  done.createDiv({ cls: "wd-eyebrow", text: `Ingested — ${pipeline.processed.length}` });
-  const list = done.createDiv({ cls: "wd-done" });
-  for (const file of pipeline.processed) {
-    const row = list.createDiv({ cls: "wd-done-row" });
-    const date = file.page.name.slice(0, 10);
-    row.createSpan({ cls: "wd-mono", text: /^\d{4}-\d{2}-\d{2}$/.test(date) ? date : "—" });
-    row.createSpan({ cls: "wd-done-name", text: file.name.replace(/\.md$/, "") });
-  }
-
-  if (pipeline.orphaned.length > 0) {
-    const orphan = pane.createDiv({ cls: "wd-block" });
-    orphan.createDiv({
-      cls: "wd-eyebrow",
-      text: `Orphaned — ${pipeline.orphaned.length}`,
-    });
-    for (const page of pipeline.orphaned) {
-      const row = orphan.createDiv({ cls: "wd-done-row" });
-      row.createSpan({ cls: "wd-done-name", text: page.name });
-      row.createSpan({ cls: "wd-mono", text: page.rawPath || "no raw path" });
-    }
-  }
-
-  return pane;
-}
-```
-
-- [ ] **Step 3: Replace the plugin class body to read the vault and render**
-
-`vault.getFiles()` returns every markdown file; filter by path prefix. `vault.cachedRead` is the correct read for display purposes.
-
-```js
-class WebinarDashPlugin extends PluginBase {
-  async onload() {
-    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
-      const cfg = parseConfig(source);
-      const root = el.createDiv({ cls: "webinar-dash" });
-      try {
-        const pipeline = await this.readPipeline(cfg);
-        renderLeftPane(root, pipeline, () => {});
-      } catch (err) {
-        root.createDiv({ cls: "wd-error", text: `Dashboard failed: ${err.message}` });
-      }
-    });
-  }
-
-  async readPipeline(cfg) {
-    const all = this.app.vault.getFiles();
-    const rawDir = cfg.rawDir.replace(/\/+$/, "") + "/";
-    const wikiDir = cfg.wikiSourceDir.replace(/\/+$/, "") + "/";
-
-    // CLAUDE.md documents raw/sources as holding "markdown/text/pdf exports".
-    // Allow-listing only "md" would hide the others from the queue with no
-    // warning — the same silent-invisibility failure this dashboard exists to
-    // remove. Wiki source pages below stay markdown-only; those really are .md.
-    const rawFiles = all
-      .filter((f) => f.path.startsWith(rawDir) && RAW_EXTENSIONS.has(f.extension))
-      .map((f) => ({ path: f.path, name: f.name, size: f.stat.size }));
-
-    const pageFiles = all.filter((f) => f.path.startsWith(wikiDir) && f.extension === "md");
-    const sourcePages = [];
-    for (const f of pageFiles) {
-      const text = await this.app.vault.cachedRead(f);
-      sourcePages.push({ path: f.path, name: f.name, rawPath: extractRawPath(text) });
-    }
-
-    return derivePipeline({ rawFiles, sourcePages });
-  }
-}
-```
-
-- [ ] **Step 4: Verify in Obsidian**
-
-Reload the app (`Ctrl+R`) and open `dashboard.md` in reading view.
-
-Expected, against the vault's current state:
-- `Queue — 3 unprocessed` listing `Agentic Engineering, explained by a 10x developer.md` (15,020 B), `Webinar Plan - From Chat Box to Your Own OS.md` (17,177 B), and `Webinar script.md` (15,841 B), each with a red **Ingest** button that does nothing yet.
-- `Ingested — 9` listing the source summaries newest first, starting with `2026-07-24`.
-- No orphaned block.
-
-- [ ] **Step 5: Run the existing tests to confirm nothing regressed**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test
-```
-
-Bare `node --test` auto-discovers the test files. Do not pass `test/` as an
-argument — Node 24 resolves a bare directory path as a module and fails with
-`MODULE_NOT_FOUND` before running anything.
-
-Expected: PASS, 9 tests.
-
-- [ ] **Step 6: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add .obsidian/plugins/webinar-dash
-git commit -m "feat: render source pipeline in left pane"
-```
-
----
-
-### Task 4: Coverage table parser
-
-**Files:**
-- Modify: `.obsidian/plugins/webinar-dash/main.js`
-- Test: `.obsidian/plugins/webinar-dash/test/coverage.test.js`
-
-**Interfaces:**
-- Consumes: `STATIONS` from Task 1
-- Produces:
-  - `parseCoverageTable(markdown) => { meta, rows, errors }`
-  - `meta`: `{ script: string | null, lastSynced: string | null }`
-  - `rows` items: `{ concept: string, status: "covered"|"partial"|"absent", stations: string[], pinned: boolean, line: number }`
-  - `errors` items: `{ line: number, text: string, reason: string }`
-  - `groupByStation(rows) => Array<{ station: string, rows: Row[] }>` ordered `All stations`, then the seven stations, then `No station`; empty groups omitted
-
-- [ ] **Step 1: Write the failing tests**
-
-Create `.obsidian/plugins/webinar-dash/test/coverage.test.js`.
-
-```js
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { parseCoverageTable, groupByStation } = require("../main.js").__test__;
-
-const DOC = [
-  "# Script coverage",
-  "",
-  "#coverage",
-  "",
-  "## Metadata",
-  "",
-  "- **Script:** `raw/sources/Webinar script.md`",
-  "- **Last synced:** 2026-07-28",
-  "",
-  "## Coverage",
-  "",
-  "| Concept | Status | Station | Pinned |",
-  "|---|---|---|---|",
-  "| [[harness]] | covered | Tools | |",
-  "| [[agentic-loops]] | partial | Process | |",
-  "| [[levels-of-ai-usage]] | partial | all | |",
-  "| [[connections-as-moat]] | absent | — | yes |",
-].join("\n");
-
-test("parseCoverageTable reads metadata", () => {
-  const { meta } = parseCoverageTable(DOC);
-  assert.equal(meta.script, "raw/sources/Webinar script.md");
-  assert.equal(meta.lastSynced, "2026-07-28");
-});
-
-test("parseCoverageTable reads every data row and skips the header", () => {
-  const { rows, errors } = parseCoverageTable(DOC);
-  assert.equal(errors.length, 0);
-  assert.equal(rows.length, 4);
-  assert.deepEqual(rows.map((r) => r.concept), [
-    "harness", "agentic-loops", "levels-of-ai-usage", "connections-as-moat",
-  ]);
-});
-
-test("parseCoverageTable normalises stations", () => {
-  const { rows } = parseCoverageTable(DOC);
-  assert.deepEqual(rows[0].stations, ["Tools"]);
-  assert.deepEqual(rows[2].stations, ["all"]);
-  assert.deepEqual(rows[3].stations, []);
-});
-
-test("parseCoverageTable reads the pinned flag", () => {
-  const { rows } = parseCoverageTable(DOC);
-  assert.equal(rows[0].pinned, false);
-  assert.equal(rows[3].pinned, true);
-});
-
-test("parseCoverageTable strips a wikilink alias", () => {
-  const doc = "| Concept | Status | Station |\n|---|---|---|\n| [[harness\\|The harness]] | covered | Tools |";
-  const { rows } = parseCoverageTable(doc);
-  assert.equal(rows[0].concept, "harness");
-});
-
-test("parseCoverageTable splits a multi-station cell", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered | Tools, Memory |";
-  const { rows } = parseCoverageTable(doc);
-  assert.deepEqual(rows[0].stations, ["Tools", "Memory"]);
-});
-
-test("parseCoverageTable rejects an invalid status instead of coercing it", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | maybe | Tools |";
-  const { rows, errors } = parseCoverageTable(doc);
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 1);
-  assert.match(errors[0].reason, /invalid status/);
-  assert.equal(errors[0].line, 3);
-});
-
-test("parseCoverageTable reports a row with too few columns", () => {
-  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered |";
-  const { rows, errors } = parseCoverageTable(doc);
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 1);
-  assert.match(errors[0].reason, /at least 3 columns/);
-});
-
-test("parseCoverageTable returns empty results for a document with no table", () => {
-  const { rows, errors } = parseCoverageTable("# Nothing\n\nJust prose.");
-  assert.equal(rows.length, 0);
-  assert.equal(errors.length, 0);
-});
-
-test("groupByStation orders all-stations first and no-station last", () => {
-  const { rows } = parseCoverageTable(DOC);
-  const groups = groupByStation(rows);
-  assert.deepEqual(groups.map((g) => g.station), [
-    "All stations", "Tools", "Process", "No station",
-  ]);
-  assert.equal(groups[1].rows[0].concept, "harness");
-});
-
-test("groupByStation places a multi-station row under its first station only", () => {
-  const rows = [{ concept: "x", status: "covered", stations: ["Memory", "Skills"], pinned: false, line: 1 }];
-  const groups = groupByStation(rows);
-  assert.equal(groups.length, 1);
-  assert.equal(groups[0].station, "Memory");
-});
-```
-
-- [ ] **Step 2: Run the tests to verify they fail**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test test/coverage.test.js
-```
-
-Expected: FAIL — `parseCoverageTable is not a function`.
-
-- [ ] **Step 3: Implement the parser**
-
-Insert into `main.js` below `derivePipeline`. The `inTable` flag flips on the `|---|` separator, which is what distinguishes the header row from data rows.
-
-```js
-const VALID_STATUS = new Set(["covered", "partial", "absent"]);
-const SEPARATOR_RE = /^\|?\s*:?-{2,}/;
-
-function parseCoverageTable(markdown) {
-  const text = String(markdown);
-  const meta = { script: null, lastSynced: null };
-
-  const scriptM = text.match(/^\s*-\s*\*\*Script:\*\*\s*`([^`]+)`/m);
-  if (scriptM) meta.script = scriptM[1].trim();
-  const syncM = text.match(/^\s*-\s*\*\*Last synced:\*\*\s*(\S+)/m);
-  if (syncM) meta.lastSynced = syncM[1].trim();
-
-  const rows = [];
-  const errors = [];
-  let inTable = false;
-
-  text.split(/\r?\n/).forEach((line, i) => {
-    const t = line.trim();
-    if (!t.startsWith("|")) {
-      inTable = false;
-      return;
-    }
-    if (SEPARATOR_RE.test(t)) {
-      inTable = true;
-      return;
-    }
-    if (!inTable) return;
-
-    const body = t.endsWith("|") ? t.slice(1, -1) : t.slice(1);
-    // Split on unescaped pipes only. Obsidian escapes the pipe of a piped
-    // wikilink inside a table cell as `\|`, and a plain split("|") tears
-    // `[[harness\|alias]]` into two cells, shifting every later column left —
-    // the status cell then reads "The harness]]" and the row is rejected.
-    const cells = body.split(/(?<!\\)\|/).map((c) => c.trim());
-
-    if (cells.length < 3) {
-      errors.push({ line: i + 1, text: t, reason: "expected at least 3 columns" });
-      return;
-    }
-
-    const status = cells[1].toLowerCase();
-    if (!VALID_STATUS.has(status)) {
-      errors.push({ line: i + 1, text: t, reason: `invalid status "${cells[1]}"` });
-      return;
-    }
-
-    // Capture up to the first ] | or backslash. Obsidian escapes the pipe in a
-    // piped wikilink inside a table cell, so the raw cell reads [[name\|alias]] —
-    // excluding the backslash is what keeps the trailing "\" out of the name.
-    const linkM = cells[0].match(/\[\[([^\]|\\]+)/);
-    const stationCell = cells[2];
-
-    rows.push({
-      concept: linkM ? linkM[1].trim() : cells[0],
-      status,
-      stations:
-        stationCell === "—" || stationCell === "-" || stationCell === ""
-          ? []
-          : stationCell.split(",").map((s) => s.trim()).filter(Boolean),
-      pinned: (cells[3] || "").toLowerCase() === "yes",
-      line: i + 1,
-    });
-  });
-
-  return { meta, rows, errors };
-}
-
-function groupByStation(rows) {
-  const order = ["All stations", ...STATIONS, "No station"];
-  const buckets = new Map(order.map((k) => [k, []]));
-
-  for (const row of rows) {
-    let key;
-    if (row.stations.includes("all")) key = "All stations";
-    else if (row.stations.length === 0) key = "No station";
-    else key = row.stations[0];
-    if (!buckets.has(key)) buckets.set(key, []);
-    buckets.get(key).push(row);
-  }
-
-  const known = order.filter((k) => buckets.get(k).length > 0);
-  const unknown = [...buckets.keys()].filter((k) => !order.includes(k) && buckets.get(k).length > 0);
-  return [...known, ...unknown].map((station) => ({ station, rows: buckets.get(station) }));
-}
-```
-
-The `\\|` in the alias test is an escaped pipe inside a JS string, which reaches the parser as a literal `|` — matching how Obsidian escapes piped wikilinks inside table cells.
-
-- [ ] **Step 4: Export them**
-
-```js
-module.exports.__test__ = {
-  parseConfig, extractRawPath, derivePipeline,
-  parseCoverageTable, groupByStation,
-  STATIONS, DEFAULTS,
-};
-```
-
-- [ ] **Step 5: Run the tests to verify they pass**
-
-```bash
-node --test test/coverage.test.js
-```
-
-Expected: PASS, 11 tests.
-
-- [ ] **Step 6: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add .obsidian/plugins/webinar-dash
-git commit -m "feat: parse the script coverage table"
-```
-
----
-
-### Task 5: Coverage data file and CLAUDE.md contract
-
-This is the wiki-side task. It creates the file the renderer reads and installs the rule that keeps it current. No plugin code changes.
-
-**Files:**
-- Create: `wiki/script-coverage.md`
-- Modify: `CLAUDE.md` (Folder Convention, Tagging Rules, Standard Workflows, Operational Commands)
-- Modify: `index.md`
-- Modify: `log.md`
-
-**Interfaces:**
-- Consumes: the row format defined in Task 4
-- Produces: `wiki/script-coverage.md` conforming to that format, with one row per page in `wiki/concepts/`
-
-- [ ] **Step 1: Create `wiki/script-coverage.md`**
-
-Statuses below are the assessment made during design, read from `raw/sources/Webinar script.md` against each concept page. All 19 concept pages are present.
-
-```markdown
-# Script coverage
-
-#coverage
-
-## Metadata
-
-- **Script:** `raw/sources/Webinar script.md`
-- **Last synced:** 2026-07-28
-- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
-
-## Coverage
-
-| Concept | Status | Station | Pinned |
-|---|---|---|---|
-| [[harness]] | covered | Tools | |
-| [[evolution-of-agent-tooling]] | covered | Tools | |
-| [[skills-as-memory]] | covered | Skills | |
-| [[solve-first-then-skillify]] | covered | Skills | |
-| [[personal-ai-operating-system]] | covered | OS | |
-| [[context-as-scarce-resource]] | partial | Memory | |
-| [[agentic-loops]] | partial | Process | |
-| [[code-as-throwaway]] | partial | OS | |
-| [[levels-of-ai-usage]] | partial | all | |
-| [[integration-dead-ends]] | absent | — | |
-| [[leave-less-room-for-imagination]] | absent | — | |
-| [[product-ownership]] | absent | — | |
-| [[connections-as-moat]] | absent | — | |
-| [[network-from-a-standing-start]] | absent | — | |
-| [[seniority-and-the-junior-squeeze]] | absent | — | |
-| [[decoupling-identity-from-profession]] | absent | — | |
-| [[think-wider-not-bigger]] | absent | — | |
-| [[make-more-cheap-code]] | absent | — | |
-| [[enterprise-ai-reality]] | absent | — | |
-
-## Notes
-
-- Eight of the ten absent concepts are human-side or strategy-side, per the grouping in `index.md`. The script's spine is machine-side and it never reaches that material.
-- The remaining two absent concepts are machine-side: [[integration-dead-ends]] and [[leave-less-room-for-imagination]]. The script demonstrates the happy path, so it never reaches connector gating or spec ambiguity — the two ways the machine side fails in practice.
-- The `ReAct` station carries no wiki concept at all.
-- Set `Pinned` to `yes` on any row whose status is a deliberate decision. Sync will not touch it.
-
-## Related Pages
-
-- [[overview]]
-- `raw/sources/Webinar script.md`
-```
-
-- [ ] **Step 2: Add the file to the folder convention in `CLAUDE.md`**
-
-In the `## Folder Convention` code block, add below the `wiki/overview.md` line:
-
-```text
-  script-coverage.md  # machine-maintained: concept coverage vs the webinar script
-```
-
-- [ ] **Step 3: Add the tag row in `CLAUDE.md`**
-
-In the Folder → tag table under `## Tagging Rules`, add a row after the `wiki/overview.md` row:
-
-```text
-| `wiki/script-coverage.md` | `#coverage`    |
-```
-
-- [ ] **Step 4: Add Workflow D to `CLAUDE.md`**
-
-Append to `## Standard Workflows`, after Workflow C:
-
-```markdown
-### Workflow D: Sync Script Coverage
-
-Maintains `wiki/script-coverage.md` — one row per page in `wiki/concepts/`, judged
-against the script named in that file's `**Script:**` metadata field.
-
-1. Read `wiki/script-coverage.md` and note every row where `Pinned` is `yes`.
-2. Read the script and every page in `wiki/concepts/`.
-3. For each concept, decide `Status` and `Station`:
-   - `covered` — the script delivers the idea, whether or not it uses the page's name.
-   - `partial` — the script gestures at it but never lands it.
-   - `absent` — the script never reaches it.
-   - `Station` is one of the seven, a comma-separated list, `all`, or `—` when absent.
-4. **Never modify a row whose `Pinned` is `yes`** — not its status, not its station.
-   A pinned row is the user's judgment and outranks yours.
-5. Add rows for concept pages with no row. Remove rows whose concept page no longer exists.
-6. Update `**Last synced:**` to today.
-7. Update `index.md` and append a `sync` entry to `log.md`.
-
-Run this workflow:
-
-- at the end of any ingest that creates or modifies a page in `wiki/concepts/`
-- whenever the script file itself changes
-- on the explicit `sync script coverage` intent
-
-The coverage baseline is always the raw script. Ingesting the script into
-`wiki/sources/` does not change the baseline.
-```
-
-- [ ] **Step 5: Add the intent to `CLAUDE.md`**
-
-In `## Operational Commands (Natural Language)`, add to the supported intents list:
-
-```markdown
-- "sync script coverage"
-```
-
-- [ ] **Step 6: Update `index.md`**
-
-Add a section after `## Sources`, and correct the stale not-yet-ingested line.
-
-```markdown
-## Coverage
-
-- [[script-coverage]] — every concept vs `raw/sources/Webinar script.md`; 5 covered, 4 partial, 10 absent
-```
-
-Replace the existing `**Raw, not yet ingested:**` line with the accurate set — the two note files live in `raw/notes/`, not `raw/sources/`, and one raw source was missing entirely:
-
-```markdown
-**Raw, not yet ingested:** `raw/sources/Agentic Engineering, explained by a 10x developer.md` · `raw/sources/Webinar Plan - From Chat Box to Your Own OS.md` · `raw/sources/Webinar script.md`
-```
-
-- [ ] **Step 7: Append to `log.md`**
-
-```markdown
-## 2026-07-28 — sync (script coverage, initial)
-- Intent: sync script coverage
-- Input: first run of Workflow D, establishing `wiki/script-coverage.md`.
-- Pages created: [[script-coverage]] (coverage) — 19 rows, one per concept page.
-- Pages updated: `CLAUDE.md` (folder convention, `#coverage` tag row, Workflow D, new intent), `index.md` (new Coverage section; corrected the not-yet-ingested list — `Ideas for webinar.md` and `my theses.md` were listed under `raw/sources/` but live in `raw/notes/`, and `Agentic Engineering, explained by a 10x developer.md` was missing entirely).
-- Notes: Initial assessment is 5 covered, 4 partial, 10 absent. Eight of the ten absent are human-side or strategy-side per this file's own concept grouping; the other two — [[integration-dead-ends]] and [[leave-less-room-for-imagination]] — are machine-side, absent because the script demos the happy path and never reaches connector gating or spec ambiguity. The `ReAct` station carries no wiki concept. No concept pages were edited.
-- Next: decide whether the human-side cluster earns a station or is a deliberate cut; decide separately whether the two absent machine-side failure modes belong in the demo; pin the rows that are decided so future syncs leave them alone.
-```
-
-- [ ] **Step 8: Verify the file parses**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node -e "
-const fs = require('fs');
-const { parseCoverageTable } = require('./main.js').__test__;
-const md = fs.readFileSync('../../../wiki/script-coverage.md', 'utf8');
-const r = parseCoverageTable(md);
-console.log('rows:', r.rows.length, 'errors:', r.errors.length);
-console.log('meta:', r.meta);
-if (r.errors.length) { console.log(r.errors); process.exit(1); }
-if (r.rows.length !== 19) { console.log('expected 19 rows'); process.exit(1); }
-"
-```
-
-Expected: `rows: 19 errors: 0` and metadata showing the script path and `2026-07-28`.
-
-- [ ] **Step 9: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add wiki/script-coverage.md CLAUDE.md index.md log.md
-git commit -m "feat: add script coverage file and Workflow D sync contract"
-```
-
----
-
-### Task 6: Right pane — coverage rendering
-
-**Files:**
-- Modify: `.obsidian/plugins/webinar-dash/main.js`
-- Modify: `.obsidian/plugins/webinar-dash/styles.css`
-
-**Interfaces:**
-- Consumes: `parseCoverageTable`, `groupByStation` from Task 4; `addIcon` from Task 3; `wiki/script-coverage.md` from Task 5
-- Produces:
-  - `reconcileConcepts(rows, conceptNames) => { rows, unsynced, stale }` where `unsynced` is concept names with no row and `stale` is rows whose concept page is gone
-  - `renderRightPane(container, parsed, reconciled)`
-
-- [ ] **Step 1: Add the reconcile helper and its tests**
-
-Append to `.obsidian/plugins/webinar-dash/test/coverage.test.js`:
-
-```js
-const { reconcileConcepts } = require("../main.js").__test__;
-
-test("reconcileConcepts finds concept pages with no row", () => {
-  const rows = [{ concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 }];
-  const out = reconcileConcepts(rows, ["harness", "brand-new-concept"]);
-  assert.deepEqual(out.unsynced, ["brand-new-concept"]);
-  assert.deepEqual(out.stale, []);
-});
-
-test("reconcileConcepts finds rows whose concept page is gone", () => {
-  const rows = [
-    { concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 },
-    { concept: "deleted-idea", status: "absent", stations: [], pinned: false, line: 2 },
-  ];
-  const out = reconcileConcepts(rows, ["harness"]);
-  assert.deepEqual(out.stale.map((r) => r.concept), ["deleted-idea"]);
-  assert.deepEqual(out.unsynced, []);
-});
-```
-
-- [ ] **Step 2: Run the tests to verify they fail**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test test/coverage.test.js
-```
-
-Expected: FAIL — `reconcileConcepts is not a function`.
-
-- [ ] **Step 3: Implement `reconcileConcepts` and export it**
-
-Insert into `main.js` below `groupByStation`:
-
-```js
-function reconcileConcepts(rows, conceptNames) {
-  const named = new Set(conceptNames);
-  const rowed = new Set(rows.map((r) => r.concept));
-  return {
-    rows,
-    unsynced: conceptNames.filter((n) => !rowed.has(n)).sort(),
-    stale: rows.filter((r) => !named.has(r.concept)),
-  };
-}
-```
-
-Add `reconcileConcepts` to the `__test__` export object.
-
-- [ ] **Step 4: Run the tests to verify they pass**
-
-```bash
-node --test test/coverage.test.js
-```
-
-Expected: PASS, 13 tests.
-
-- [ ] **Step 5: Add the coverage styles**
-
-Append to `styles.css`:
-
-```css
-.wd-meter { display: flex; gap: 2px; height: 8px; width: 100%; }
-.wd-meter > i { display: block; height: 100%; }
-.wd-seg-covered { background: var(--wd-ok); }
-.wd-seg-partial { background: var(--wd-warn); }
-.wd-seg-absent { background: var(--wd-danger); }
-
-.wd-key { display: flex; flex-wrap: wrap; gap: 16px; font-size: 12px; color: var(--wd-fg-2); }
-.wd-key > span { display: inline-flex; align-items: center; gap: 6px; }
-.wd-key i { width: 8px; height: 8px; flex: none; }
-
-.wd-tbl { width: 100%; border-collapse: collapse; font-size: 13px; }
-.wd-tbl th {
-  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
-  text-transform: uppercase; color: var(--wd-fg-3); font-weight: 500;
-  text-align: left; padding: 0 12px 8px 0;
-  border-bottom: 1px solid var(--wd-border-strong);
-}
-.wd-tbl td {
-  padding: 6px 12px 6px 0; border-bottom: 1px solid var(--wd-border-subtle);
-  vertical-align: baseline;
-}
-.wd-grp td {
-  padding-top: 16px; border-bottom: 1px solid var(--wd-border-strong);
-  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
-  text-transform: uppercase; color: var(--wd-fg); font-weight: 500;
-}
-.wd-stat { display: inline-flex; align-items: center; gap: 5px; font-size: 12px; font-weight: 600; }
-.wd-stat svg { width: 13px; height: 13px; flex: none; }
-.wd-covered { color: var(--wd-ok); }
-.wd-partial { color: var(--wd-warn); }
-.wd-absent { color: var(--wd-danger); }
-.wd-pin { font-family: var(--wd-mono); font-size: 10px; color: var(--wd-fg-4); }
-```
-
-- [ ] **Step 6: Add the right-pane renderer**
-
-Insert into `main.js` below `renderLeftPane`:
-
-```js
-const STATUS_ICON = { covered: "check", partial: "minus", absent: "x" };
-
-function renderRightPane(container, parsed, reconciled) {
-  const pane = container.createDiv({ cls: "wd-pane" });
-  pane.createDiv({ cls: "wd-eyebrow", text: "Coverage — by station" });
-
-  const counts = { covered: 0, partial: 0, absent: 0 };
-  for (const row of parsed.rows) counts[row.status] += 1;
-  const total = parsed.rows.length;
-
-  if (total > 0) {
-    const meter = pane.createDiv({ cls: "wd-meter" });
-    meter.setAttr(
-      "aria-label",
-      `Coverage: ${counts.covered} covered, ${counts.partial} partial, ${counts.absent} absent of ${total}`
-    );
-    for (const key of ["covered", "partial", "absent"]) {
-      if (counts[key] === 0) continue;
-      const seg = meter.createEl("i", { cls: `wd-seg-${key}` });
-      seg.style.flex = String(counts[key]);
-    }
-    const key = pane.createDiv({ cls: "wd-key" });
-    for (const name of ["covered", "partial", "absent"]) {
-      const span = key.createSpan();
-      span.createEl("i", { cls: `wd-seg-${name}` });
-      span.createSpan({ text: `${counts[name]} ${name}` });
-    }
-  }
-
-  if (parsed.errors.length > 0) {
-    const box = pane.createDiv({ cls: "wd-error" });
-    box.createDiv({ text: `${parsed.errors.length} unparseable row(s):` });
-    for (const e of parsed.errors) {
-      box.createDiv({ cls: "wd-mono", text: `line ${e.line} — ${e.reason}` });
-    }
-  }
-
-  const table = pane.createEl("table", { cls: "wd-tbl" });
-  const head = table.createEl("thead").createEl("tr");
-  for (const h of ["Concept", "Status", "Station", ""]) head.createEl("th", { text: h });
-  const body = table.createEl("tbody");
-
-  for (const group of groupByStation(parsed.rows)) {
-    const gr = body.createEl("tr", { cls: "wd-grp" });
-    gr.createEl("td", { attr: { colspan: "4" }, text: `${group.station} — ${group.rows.length}` });
-    for (const row of group.rows) {
-      const tr = body.createEl("tr");
-      // Obsidian's click handler resolves internal links via data-href, so both
-      // attributes are required for the link to open the concept page.
-      tr.createEl("td").createEl("a", {
-        cls: "internal-link",
-        text: row.concept,
-        attr: { href: row.concept, "data-href": row.concept },
-      });
-      const stat = tr.createEl("td").createSpan({ cls: `wd-stat wd-${row.status}` });
-      addIcon(stat, STATUS_ICON[row.status]);
-      stat.createSpan({ text: row.status });
-      tr.createEl("td", { cls: "wd-mono", text: row.stations.join(", ") || "—" });
-      tr.createEl("td", { cls: "wd-pin", text: row.pinned ? "pinned" : "" });
-    }
-  }
-
-  if (reconciled.unsynced.length > 0) {
-    const box = pane.createDiv({ cls: "wd-block" });
-    box.createDiv({ cls: "wd-eyebrow", text: `Unsynced — ${reconciled.unsynced.length}` });
-    box.createDiv({
-      cls: "wd-note",
-      text: `Concept pages with no row. Run "sync script coverage": ${reconciled.unsynced.join(", ")}`,
-    });
-  }
-
-  if (reconciled.stale.length > 0) {
-    const box = pane.createDiv({ cls: "wd-block" });
-    box.createDiv({ cls: "wd-eyebrow", text: `Stale — ${reconciled.stale.length}` });
-    box.createDiv({
-      cls: "wd-note",
-      text: `Rows whose concept page is gone: ${reconciled.stale.map((r) => r.concept).join(", ")}`,
-    });
-  }
-
-  return pane;
-}
-```
-
-- [ ] **Step 7: Wire it into the plugin class**
-
-Replace the code-block processor callback and add a reader method:
-
-```js
-    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
-      const cfg = parseConfig(source);
-      const root = el.createDiv({ cls: "webinar-dash" });
-      try {
-        const pipeline = await this.readPipeline(cfg);
-        renderLeftPane(root, pipeline, () => {});
-      } catch (err) {
-        root.createDiv({ cls: "wd-error", text: `Pipeline failed: ${err.message}` });
-      }
-      try {
-        const { parsed, reconciled } = await this.readCoverage(cfg);
-        renderRightPane(root, parsed, reconciled);
-      } catch (err) {
-        root.createDiv({ cls: "wd-error", text: `Coverage failed: ${err.message}` });
-      }
-    });
-```
-
-The two panes render in separate `try` blocks so a broken coverage file never blanks the source pipeline.
-
-```js
-  async readCoverage(cfg) {
-    const file = this.app.vault.getAbstractFileByPath(cfg.coverage);
-    if (!file) throw new Error(`no coverage file at ${cfg.coverage}`);
-    const parsed = parseCoverageTable(await this.app.vault.cachedRead(file));
-
-    const conceptDir = cfg.conceptDir.replace(/\/+$/, "") + "/";
-    const conceptNames = this.app.vault
-      .getFiles()
-      .filter((f) => f.path.startsWith(conceptDir) && f.extension === "md")
-      .map((f) => f.basename);
-
-    return { parsed, reconciled: reconcileConcepts(parsed.rows, conceptNames) };
-  }
-```
-
-- [ ] **Step 8: Verify in Obsidian**
-
-Reload (`Ctrl+R`) and open `dashboard.md`.
-
-Expected: a right pane with a meter reading `5 covered`, `4 partial`, `10 absent`; a table grouped `All stations — 1`, `Tools — 2`, `Memory — 1`, `Skills — 2`, `Process — 1`, `OS — 2`, `No station — 10`; no unsynced block; no stale block; concept names clickable through to their pages.
-
-- [ ] **Step 9: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add .obsidian/plugins/webinar-dash
-git commit -m "feat: render coverage meter and station-grouped table"
-```
-
----
-
-### Task 7: Headless ingest
-
-**Files:**
-- Modify: `.obsidian/plugins/webinar-dash/main.js`
-- Modify: `.obsidian/plugins/webinar-dash/styles.css`
-- Test: `.obsidian/plugins/webinar-dash/test/safety.test.js`
-
-**Interfaces:**
-- Consumes: `renderLeftPane`'s `onIngest(file, rowEl)` callback from Task 3
-- Produces:
-  - `isSafeFilename(name) => boolean`
-  - `runIngest(file, rowEl)` as a plugin method
-
-- [ ] **Step 1: Write the failing tests**
-
-Create `.obsidian/plugins/webinar-dash/test/safety.test.js`. The accept cases are real filenames from this vault.
-
-```js
-"use strict";
-const test = require("node:test");
-const assert = require("node:assert/strict");
-const { isSafeFilename } = require("../main.js").__test__;
-
-test("isSafeFilename accepts every filename currently in the vault", () => {
-  const real = [
-    "Agentic Engineering, explained by a 10x developer.md",
-    "Webinar Plan - From Chat Box to Your Own OS.md",
-    "Webinar script.md",
-    "You're reading way too much code.md",
-    "ИИ глупый!.md",
-    "Скиллы на базе git — новая память AI-агентов.md",
-    "sebastian interview - conclusions and insights.md",
-    "In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md",
-  ];
-  for (const name of real) {
-    assert.equal(isSafeFilename(name), true, `should accept: ${name}`);
-  }
-});
-
-test("isSafeFilename rejects shell metacharacters", () => {
-  for (const bad of ['a".md', "a`b.md", "a$b.md", "a&b.md", "a|b.md", "a;b.md",
-                     "a<b.md", "a>b.md", "a%b.md", "a\nb.md", "a\rb.md"]) {
-    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
-  }
-});
-
-test("isSafeFilename rejects the cmd.exe escape character and control characters", () => {
-  // `^` escapes the next character in cmd.exe, so it can defuse the closing
-  // quote. NUL additionally makes spawn() throw synchronously.
-  for (const bad of ["a^b.md", "a\u0000b.md", "a\u001bb.md", "a\u007fb.md"]) {
-    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
-  }
-});
-
-test("isSafeFilename rejects path traversal", () => {
-  assert.equal(isSafeFilename("../secrets.md"), false);
-  assert.equal(isSafeFilename("a/../../b.md"), false);
-});
-
-test("isSafeFilename rejects empty and non-string input", () => {
-  assert.equal(isSafeFilename(""), false);
-  assert.equal(isSafeFilename(null), false);
-  assert.equal(isSafeFilename(undefined), false);
-  assert.equal(isSafeFilename(42), false);
-});
-```
-
-- [ ] **Step 2: Run the tests to verify they fail**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test test/safety.test.js
-```
-
-Expected: FAIL — `isSafeFilename is not a function`.
-
-- [ ] **Step 3: Implement the validator and export it**
-
-Insert into `main.js` below `reconcileConcepts`:
-
-```js
-// `shell: true` is required on Windows to resolve `claude.cmd`, which puts the
-// filename into a shell string. Reject anything cmd.exe or a POSIX shell would
-// interpret, plus every control character.
-//
-// `^` is cmd.exe's escape character and belongs to the same class Node escapes
-// in its CVE-2024-27980 mitigation for exactly this spawn-through-.cmd shape.
-// Control characters are rejected because a NUL byte makes spawn() throw
-// synchronously, which would otherwise strand the row mid-run.
-//
-// Apostrophes, spaces, cyrillic, em dashes and `!` stay allowed — they are safe
-// inside double quotes and appear in real filenames in this vault. (`!` would
-// matter only under `setlocal enabledelayedexpansion`, which is not in play.)
-const UNSAFE_CHARS = /["`$&|;<>%^\u0000-\u001f\u007f]/;
-
-function isSafeFilename(name) {
-  if (typeof name !== "string" || name.length === 0) return false;
-  if (UNSAFE_CHARS.test(name)) return false;
-  if (name.includes("..")) return false;
-  return true;
-}
-```
-
-Add `isSafeFilename` to the `__test__` export object.
-
-- [ ] **Step 4: Run the tests to verify they pass**
-
-```bash
-node --test test/safety.test.js
-```
-
-Expected: PASS, 5 tests.
-
-- [ ] **Step 5: Add ingest-state styles**
-
-Append to `styles.css`:
-
-```css
-.wd-status { font-family: var(--wd-mono); font-size: 11px; flex: none; }
-.wd-status-running { color: var(--wd-warn); }
-.wd-status-done { color: var(--wd-ok); }
-.wd-status-failed { color: var(--wd-danger); }
-.wd-output {
-  font-family: var(--wd-mono); font-size: 11px; white-space: pre-wrap;
-  max-height: 220px; overflow: auto; margin-top: 8px;
-  border: 1px solid var(--wd-border); border-radius: 4px; padding: 8px;
-  color: var(--wd-fg-2);
-}
-```
-
-- [ ] **Step 6: Implement `runIngest` on the plugin class**
-
-`getBasePath()` exists on `FileSystemAdapter`; on mobile the adapter is a different class and `child_process` is unavailable, which is why the manifest sets `isDesktopOnly`.
-
-```js
-Add the in-flight registry as the first line of `onload()`, before the code-block
-processor is registered. It lives on the plugin instance so it survives the row
-re-renders that a DOM-scoped guard cannot:
-
-```js
-  async onload() {
-    this.running = new Set();
-    // ... existing registerMarkdownCodeBlockProcessor call follows unchanged
-```
-
-Then add both methods to the class:
-
-```js
-  vaultPath() {
-    const adapter = this.app.vault.adapter;
-    if (typeof adapter.getBasePath === "function") return adapter.getBasePath();
-    return null;
-  }
-
-  runIngest(file, rowEl) {
-    const Notice = OB ? OB.Notice : null;
-    const notify = (msg) => { if (Notice) new Notice(msg); };
-
-    // In-flight state is keyed on the file, not the row. Obsidian rebuilds the
-    // row on every re-render, so a row-scoped guard would let a re-render hand
-    // out a fresh row whose guard is unset — and a second click would then run
-    // a second unsupervised agent against the same file, concurrently writing
-    // the same wiki pages as the first.
-    if (this.running.has(file.path)) {
-      notify(`Already ingesting ${file.name}.`);
-      return;
-    }
-
-    if (!isSafeFilename(file.name)) {
-      rowEl.createDiv({
-        cls: "wd-output",
-        text: `Refused: "${file.name}" contains a character that is unsafe to pass to a shell.`,
-      });
-      return;
-    }
-
-    const base = this.vaultPath();
-    if (!base) {
-      notify("Ingest needs desktop Obsidian.");
-      return;
-    }
-
-    let spawn;
-    try {
-      ({ spawn } = require("child_process"));
-    } catch (_) {
-      notify("child_process unavailable — ingest needs desktop Obsidian.");
-      return;
-    }
-
-    // A retry reuses the same row. Clear the previous run's status and output
-    // so they are replaced rather than stacked on top of each other.
-    rowEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
-
-    const button = rowEl.querySelector("button");
-    if (button) button.disabled = true;
-    this.running.add(file.path);
-
-    const status = rowEl.createSpan({ cls: "wd-status wd-status-running", text: "running 0s" });
-    const output = rowEl.createDiv({ cls: "wd-output", text: "" });
-    const started = Date.now();
-    const timer = window.setInterval(() => {
-      status.setText(`running ${Math.round((Date.now() - started) / 1000)}s`);
-    }, 1000);
-    this.registerInterval(timer);
-
-    // Bounded as it accumulates, not only when displayed, so a long or noisy
-    // run cannot grow this string without limit.
-    let buffered = "";
-    const append = (text) => {
-      buffered = (buffered + text).slice(-8000);
-      output.setText(buffered);
-      output.scrollTop = output.scrollHeight;
-    };
-
-    // Single exit path: every way this run can end clears the timer, releases
-    // the file, and re-enables the button.
-    const finish = (cls, label) => {
-      window.clearInterval(timer);
-      this.running.delete(file.path);
-      status.className = `wd-status ${cls}`;
-      status.setText(label);
-      if (button) button.disabled = false;
-    };
-
-    let child;
-    try {
-      child = spawn("claude", ["-p", `ingest "${file.name}"`], {
-        cwd: base,
-        shell: true,
-      });
-    } catch (err) {
-      // spawn() throws synchronously for some argument shapes. Without this the
-      // timer would run forever and the row would stay disabled until reload.
-      append(`\nCould not start claude: ${err.message}`);
-      finish("wd-status-failed", "failed");
-      return;
-    }
-
-    // Decode as UTF-8 across chunk boundaries. Raw Buffer chunks split wherever
-    // the OS buffer ends, and this vault's output is full of Cyrillic and em
-    // dashes that would otherwise decode as replacement characters.
-    child.stdout.setEncoding("utf8");
-    child.stderr.setEncoding("utf8");
-    child.stdout.on("data", append);
-    child.stderr.on("data", append);
-
-    child.on("error", (err) => {
-      append(`\nCould not start claude: ${err.message}\nIs it on PATH?`);
-      finish("wd-status-failed", "failed");
-    });
-
-    child.on("close", (code) => {
-      const secs = Math.round((Date.now() - started) / 1000);
-      if (code === 0) {
-        finish("wd-status-done", `done in ${secs}s`);
-        notify(`Ingested ${file.name}. Reopen the dashboard to refresh.`);
-      } else {
-        finish("wd-status-failed", `failed - exit ${code}`);
-      }
-    });
-  }
-```
-
-- [ ] **Step 7: Pass the real handler into the left pane**
-
-In the code-block processor, replace the no-op:
-
-```js
-        renderLeftPane(root, pipeline, (file, rowEl) => this.runIngest(file, rowEl));
-```
-
-- [ ] **Step 8: Verify the refusal path without spawning anything**
-
-Temporarily rename a raw file to include a shell metacharacter, reload, and click Ingest.
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/raw/sources"
-cp "Webinar script.md" 'bad&name.md'
-```
-
-Expected in Obsidian: the row shows `Refused: "bad&name.md" contains a character that is unsafe to pass to a shell.` and no process starts.
-
-Then remove it:
-
-```bash
-rm 'bad&name.md'
-```
-
-- [ ] **Step 9: Verify a real ingest**
-
-Reload Obsidian and click **Ingest** on `Agentic Engineering, explained by a 10x developer.md`.
-
-Expected: the button disables, the status ticks `running 1s`, `running 2s`, …, streamed output appears below the row, and on completion the status reads `done in Ns`. Reopen `dashboard.md`: the queue drops to 2 and the ingested list rises to 10.
-
-This writes to the vault unsupervised. Everything is committed, so `git diff HEAD` shows exactly what the agent changed and `git checkout -- .` reverts it.
-
-- [ ] **Step 10: Run the whole suite**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
-node --test
-```
-
-Bare `node --test` auto-discovers the test files. Do not pass `test/` as an
-argument — Node 24 resolves a bare directory path as a module and fails with
-`MODULE_NOT_FOUND` before running anything.
-
-Expected: PASS, 27 tests across three files.
-
-- [ ] **Step 11: Commit**
-
-```bash
-cd "D:/Projects/Notes/Webinar/Webinar"
-git add .obsidian/plugins/webinar-dash
-git commit -m "feat: spawn headless claude ingest with filename validation"
-```
-
----
-
-## Self-Review
-
-**Spec coverage.** Every section of the spec maps to a task:
-
-| Spec section | Task |
-|---|---|
-| Architecture, config block, no build step | 1 |
-| Source pipeline derivation, `**Raw path:**` join | 2 |
-| Left pane, tesanti styling, dark mode | 3 |
-| Coverage file format, parsing rules | 4 |
-| Coverage data file, four `CLAUDE.md` changes, `index.md`, `log.md` | 5 |
-| Two-pane layout, meter, station grouping, `unsynced` / `stale` | 6 |
-| Ingest spawn, filename rejection list, per-row states, preflight | 7 |
-| Failure modes table | 3 (orphaned), 6 (parse errors, unsynced, stale), 7 (no adapter, no `claude`) |
-
-The spec's "v1 does not write the coverage file from the plugin" is honored — no task adds a pin toggle.
-
-**Placeholder scan.** No `TBD`, no "add error handling", no "similar to Task N". Every code step carries runnable code.
-
-**Type consistency.** `rawFiles` items are `{path, name, size}` in Tasks 2 and 3. `sourcePages` items are `{path, name, rawPath}` in both. Coverage rows are `{concept, status, stations, pinned, line}` in Tasks 4, 5, and 6. `onIngest(file, rowEl)` is declared in Task 3 and implemented with the same signature in Task 7. `addIcon(parent, name)` is defined in Task 3 and reused in Task 6. `STATIONS` is defined in Task 1 and consumed by `groupByStation` in Task 4.
-
-**Test count.** Task 2 adds 9, Task 4 adds 11, Task 6 adds 2, Task 7 adds 5 — 27 total, matching Step 10 of Task 7.
+# Webinar Vault Dashboard Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Build an Obsidian plugin that renders a two-pane dashboard showing the vault's source pipeline (with one-click headless ingest) and concept coverage against the webinar script.
+
+**Architecture:** A single unbundled `main.js` CommonJS plugin. Pure logic (path extraction, pipeline derivation, table parsing, filename validation) lives at the top of the file behind a guarded `require("obsidian")`, so it can be unit-tested with plain `node --test` and no Obsidian runtime. Obsidian glue (code-block processor, DOM rendering, subprocess spawn) sits below it. The plugin reads the vault and never writes to it.
+
+**Tech Stack:** Plain CommonJS (no TypeScript, no bundler, no build step), Node's built-in `node:test` runner, Node `child_process`, Obsidian Plugin API, tesanti design tokens.
+
+## Global Constraints
+
+- **No build step.** `main.js` is loaded by Obsidian verbatim. No TypeScript, no esbuild, no `npm install`.
+- **No dependencies.** Tests use Node's built-in `node:test` and `node:assert/strict` only.
+- **`isDesktopOnly: true`** is mandatory in `manifest.json` — the plugin uses `child_process`, which Obsidian only provides on desktop.
+- **`minAppVersion`: `"1.5.0"`.**
+- **Relative `require()` between plugin files is unsupported.** All code lives in one `main.js`.
+- **The plugin never writes to the vault.** It reads files and spawns one subprocess. It must never read `index.md`.
+- **Design tokens are copied verbatim** from `tesanti Design System.zip → colors_and_type.css`. Black / white / red only; radii at most 6px; 1px hairlines instead of shadows.
+- **No emoji anywhere** — in code, UI, comments, or commit messages. Status icons are inline Lucide stroke paths.
+- **Sentence case** for all UI copy. No Title Case.
+- **The seven stations**, in order: `Chat box`, `ReAct`, `Tools`, `Memory`, `Skills`, `Process`, `OS`.
+- **Valid statuses**, exactly: `covered`, `partial`, `absent`.
+
+---
+
+## File Structure
+
+| File | Responsibility |
+|---|---|
+| `.obsidian/plugins/webinar-dash/manifest.json` | Plugin identity and desktop-only flag |
+| `.obsidian/plugins/webinar-dash/main.js` | Everything: pure helpers, renderers, plugin class |
+| `.obsidian/plugins/webinar-dash/styles.css` | tesanti tokens and dashboard styling, scoped to `.webinar-dash` |
+| `.obsidian/plugins/webinar-dash/test/pipeline.test.js` | Tests for `extractRawPath`, `derivePipeline` |
+| `.obsidian/plugins/webinar-dash/test/coverage.test.js` | Tests for `parseCoverageTable`, `groupByStation` |
+| `.obsidian/plugins/webinar-dash/test/safety.test.js` | Tests for `isSafeFilename` |
+| `dashboard.md` | Vault root. Holds one `webinar-dash` config block |
+| `wiki/script-coverage.md` | The coverage table |
+| `CLAUDE.md` | Modified: folder convention, tagging rules, Workflow D, sync triggers |
+
+Tests live inside the plugin folder. Obsidian loads only `main.js` from a plugin directory, so `test/` is inert at runtime.
+
+---
+
+### Task 1: Repository and plugin scaffold
+
+**Files:**
+- Create: `.gitignore`
+- Create: `.obsidian/plugins/webinar-dash/manifest.json`
+- Create: `.obsidian/plugins/webinar-dash/main.js`
+- Create: `dashboard.md`
+
+**Interfaces:**
+- Consumes: nothing
+- Produces: a loadable plugin registering the `webinar-dash` code-block language; `module.exports.__test__` as the export surface every later task extends
+
+- [ ] **Step 1: Initialize the repository**
+
+The vault is not currently a git repository. The plugin will spawn agents that write to `wiki/`, `index.md`, and `log.md` unsupervised, so an undo path is required before that capability exists.
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git init
+```
+
+- [ ] **Step 2: Create `.gitignore`**
+
+Obsidian rewrites `workspace.json` constantly; it is local UI state, not vault content.
+
+```gitignore
+.obsidian/workspace.json
+.obsidian/workspace-mobile.json
+.obsidian/cache
+```
+
+- [ ] **Step 3: Commit the vault as it stands**
+
+```bash
+git add -A
+git commit -m "chore: initial commit of vault before dashboard work"
+```
+
+- [ ] **Step 4: Create the manifest**
+
+`isDesktopOnly` must be `true` — Obsidian only exposes `child_process` on desktop, and the ingest button depends on it.
+
+```json
+{
+  "id": "webinar-dash",
+  "name": "Webinar dashboard",
+  "version": "0.1.0",
+  "minAppVersion": "1.5.0",
+  "description": "Source pipeline and script coverage for the webinar vault.",
+  "author": "meels",
+  "isDesktopOnly": true
+}
+```
+
+- [ ] **Step 5: Create `main.js` with the Obsidian guard and a stub renderer**
+
+The `try/catch` around `require("obsidian")` is what lets `node --test` load this file. Under Node the require throws, `OB` stays null, and `PluginBase` becomes an empty class so `class ... extends PluginBase` still evaluates. Every later task adds pure functions above the plugin class and registers them in `__test__`.
+
+```js
+"use strict";
+
+// Obsidian injects its own module resolver. Under plain `node --test` it is
+// absent, so guard the require and fall back to an empty base class. This is
+// what keeps the pure helpers below unit-testable without an Obsidian runtime.
+let OB = null;
+try {
+  OB = require("obsidian");
+} catch (_) {
+  OB = null;
+}
+const PluginBase = OB ? OB.Plugin : class {};
+
+const STATIONS = ["Chat box", "ReAct", "Tools", "Memory", "Skills", "Process", "OS"];
+
+const DEFAULTS = {
+  script: "raw/sources/Webinar script.md",
+  coverage: "wiki/script-coverage.md",
+  rawDir: "raw/sources",
+  wikiSourceDir: "wiki/sources",
+  conceptDir: "wiki/concepts",
+};
+
+function parseConfig(source) {
+  const cfg = Object.assign({}, DEFAULTS);
+  for (const line of String(source).split(/\r?\n/)) {
+    const m = line.match(/^\s*([A-Za-z][A-Za-z0-9_]*)\s*:\s*(.+?)\s*$/);
+    if (m && Object.prototype.hasOwnProperty.call(DEFAULTS, m[1])) {
+      cfg[m[1]] = m[2];
+    }
+  }
+  return cfg;
+}
+
+class WebinarDashPlugin extends PluginBase {
+  async onload() {
+    this.registerMarkdownCodeBlockProcessor("webinar-dash", (source, el, ctx) => {
+      const cfg = parseConfig(source);
+      const root = el.createDiv({ cls: "webinar-dash" });
+      root.createDiv({ cls: "wd-eyebrow", text: "Webinar dashboard" });
+      root.createEl("p", { text: `Reading coverage from ${cfg.coverage}` });
+    });
+  }
+}
+
+module.exports = WebinarDashPlugin;
+module.exports.default = WebinarDashPlugin;
+module.exports.__test__ = { parseConfig, STATIONS, DEFAULTS };
+```
+
+- [ ] **Step 6: Create `dashboard.md`**
+
+````markdown
+# Webinar dashboard
+
+```webinar-dash
+script: raw/sources/Webinar script.md
+coverage: wiki/script-coverage.md
+```
+````
+
+- [ ] **Step 7: Enable the plugin and verify it renders**
+
+In Obsidian: Settings → Community plugins → turn off Restricted mode if on → Installed plugins → Reload → enable "Webinar dashboard". Open `dashboard.md` in reading view.
+
+Expected: the text `Webinar dashboard` followed by `Reading coverage from wiki/script-coverage.md`.
+
+If nothing renders, run "Reload app without saving" (`Ctrl+R`) — Obsidian caches plugin code between edits.
+
+- [ ] **Step 8: Commit**
+
+```bash
+git add .gitignore .obsidian/plugins/webinar-dash dashboard.md
+git commit -m "feat: scaffold webinar-dash plugin with config block"
+```
+
+---
+
+### Task 2: Source pipeline derivation
+
+**Files:**
+- Modify: `.obsidian/plugins/webinar-dash/main.js`
+- Test: `.obsidian/plugins/webinar-dash/test/pipeline.test.js`
+
+**Interfaces:**
+- Consumes: `module.exports.__test__` from Task 1
+- Produces:
+  - `extractRawPath(text: string) => string | null`
+  - `derivePipeline({ rawFiles, sourcePages }) => { processed, unprocessed, orphaned }`
+  - `rawFiles` items: `{ path: string, name: string, size: number }`
+  - `sourcePages` items: `{ path: string, name: string, rawPath: string | null }`
+  - `processed` items: `rawFiles` item plus `{ page }`
+
+- [ ] **Step 1: Write the failing tests**
+
+Create `.obsidian/plugins/webinar-dash/test/pipeline.test.js`. The fixtures are the real strings from this vault.
+
+```js
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { extractRawPath, derivePipeline } = require("../main.js").__test__;
+
+test("extractRawPath pulls the backticked path", () => {
+  const page = [
+    "# You're reading way too much code",
+    "",
+    "#source",
+    "",
+    "## Source Metadata",
+    "",
+    "- **Date:** YouTube video, 24:11",
+    "- **Raw path:** `raw/sources/You're reading way too much code.md`",
+    "- **Source type:** video essay",
+  ].join("\n");
+  assert.equal(extractRawPath(page), "raw/sources/You're reading way too much code.md");
+});
+
+test("extractRawPath handles cyrillic and em dashes", () => {
+  const page = "- **Raw path:** `raw/sources/Скиллы на базе git — новая память AI-агентов.md`";
+  assert.equal(extractRawPath(page), "raw/sources/Скиллы на базе git — новая память AI-агентов.md");
+});
+
+test("extractRawPath returns null when the line is absent", () => {
+  assert.equal(extractRawPath("# A page\n\n#source\n\nNo metadata here."), null);
+});
+
+test("derivePipeline splits claimed from unclaimed raw files", () => {
+  const rawFiles = [
+    { path: "raw/sources/Nina interview.md", name: "Nina interview.md", size: 6614 },
+    { path: "raw/sources/Webinar script.md", name: "Webinar script.md", size: 15841 },
+  ];
+  const sourcePages = [
+    {
+      path: "wiki/sources/2026-07-14-nina-interview.md",
+      name: "2026-07-14-nina-interview.md",
+      rawPath: "raw/sources/Nina interview.md",
+    },
+  ];
+  const out = derivePipeline({ rawFiles, sourcePages });
+  assert.equal(out.processed.length, 1);
+  assert.equal(out.processed[0].name, "Nina interview.md");
+  assert.equal(out.processed[0].page.name, "2026-07-14-nina-interview.md");
+  assert.equal(out.unprocessed.length, 1);
+  assert.equal(out.unprocessed[0].name, "Webinar script.md");
+  assert.equal(out.orphaned.length, 0);
+});
+
+test("derivePipeline reports source pages whose raw file is gone", () => {
+  const out = derivePipeline({
+    rawFiles: [],
+    sourcePages: [
+      { path: "wiki/sources/x.md", name: "x.md", rawPath: "raw/sources/deleted.md" },
+    ],
+  });
+  assert.equal(out.orphaned.length, 1);
+  assert.equal(out.orphaned[0].name, "x.md");
+});
+
+test("derivePipeline treats a page with no raw path as orphaned", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
+    sourcePages: [{ path: "wiki/sources/y.md", name: "y.md", rawPath: null }],
+  });
+  assert.equal(out.orphaned.length, 1);
+  assert.equal(out.unprocessed.length, 1);
+});
+
+test("derivePipeline routes a duplicate raw-path claim to orphaned", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/a.md", name: "a.md", size: 10 }],
+    sourcePages: [
+      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/a.md" },
+      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/a.md" },
+    ],
+  });
+  assert.equal(out.processed.length, 1);
+  assert.equal(out.processed[0].page.name, "first.md");
+  assert.equal(out.unprocessed.length, 0);
+  assert.deepEqual(out.orphaned.map((p) => p.name), ["second.md"]);
+});
+
+test("derivePipeline lists a page once when it is both a duplicate claim and missing its raw file", () => {
+  const out = derivePipeline({
+    rawFiles: [{ path: "raw/sources/other.md", name: "other.md", size: 10 }],
+    sourcePages: [
+      { path: "wiki/sources/first.md", name: "first.md", rawPath: "raw/sources/gone.md" },
+      { path: "wiki/sources/second.md", name: "second.md", rawPath: "raw/sources/gone.md" },
+    ],
+  });
+  assert.deepEqual(out.orphaned.map((p) => p.name), ["first.md", "second.md"]);
+});
+
+test("derivePipeline sorts unprocessed by name and processed newest first", () => {
+  const out = derivePipeline({
+    rawFiles: [
+      { path: "raw/sources/b.md", name: "b.md", size: 1 },
+      { path: "raw/sources/a.md", name: "a.md", size: 1 },
+      { path: "raw/sources/c.md", name: "c.md", size: 1 },
+    ],
+    sourcePages: [
+      { path: "wiki/sources/2026-07-14-x.md", name: "2026-07-14-x.md", rawPath: "raw/sources/c.md" },
+    ],
+  });
+  assert.deepEqual(out.unprocessed.map((f) => f.name), ["a.md", "b.md"]);
+  assert.equal(out.processed.length, 1);
+});
+```
+
+- [ ] **Step 2: Run the tests to verify they fail**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test test/pipeline.test.js
+```
+
+Expected: FAIL — `extractRawPath is not a function` (it is not yet exported).
+
+- [ ] **Step 3: Implement the helpers**
+
+In `main.js`, insert directly below the `DEFAULTS` constant:
+
+```js
+const RAW_PATH_RE = /^\s*-\s*\*\*Raw path:\*\*\s*`([^`]+)`/m;
+
+function extractRawPath(text) {
+  const m = String(text).match(RAW_PATH_RE);
+  return m ? m[1].trim() : null;
+}
+
+function derivePipeline({ rawFiles, sourcePages }) {
+  // First claim on a raw path wins. A later page claiming the same file is a
+  // duplicate claim — real catalog drift — and joins `orphaned` rather than
+  // being silently dropped. `orphaned` therefore means "source page not paired
+  // with a raw file", whatever the reason.
+  const claimed = new Map();
+  const duplicates = [];
+  for (const page of sourcePages) {
+    if (!page.rawPath) continue;
+    if (claimed.has(page.rawPath)) duplicates.push(page);
+    else claimed.set(page.rawPath, page);
+  }
+
+  const processed = [];
+  const unprocessed = [];
+  for (const file of rawFiles) {
+    const page = claimed.get(file.path);
+    if (page) processed.push(Object.assign({}, file, { page }));
+    else unprocessed.push(file);
+  }
+
+  // One pass over sourcePages, so a page appears in `orphaned` at most once no
+  // matter how many of the three reasons apply to it. Concatenating a separate
+  // duplicates array here would double-count a losing claimant whose shared raw
+  // path is also missing from disk.
+  const rawPaths = new Set(rawFiles.map((f) => f.path));
+  const duplicateSet = new Set(duplicates);
+  const orphaned = sourcePages.filter(
+    (p) => duplicateSet.has(p) || !p.rawPath || !rawPaths.has(p.rawPath)
+  );
+
+  unprocessed.sort((a, b) => a.name.localeCompare(b.name));
+  processed.sort((a, b) => b.page.name.localeCompare(a.page.name));
+
+  return { processed, unprocessed, orphaned };
+}
+```
+
+- [ ] **Step 4: Export them**
+
+Replace the `__test__` line at the bottom of `main.js`:
+
+```js
+module.exports.__test__ = { parseConfig, extractRawPath, derivePipeline, STATIONS, DEFAULTS };
+```
+
+- [ ] **Step 5: Run the tests to verify they pass**
+
+```bash
+node --test test/pipeline.test.js
+```
+
+Expected: PASS, 9 tests.
+
+- [ ] **Step 6: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add .obsidian/plugins/webinar-dash
+git commit -m "feat: derive source pipeline from raw path claims"
+```
+
+---
+
+### Task 3: Left pane — source pipeline rendering
+
+**Files:**
+- Modify: `.obsidian/plugins/webinar-dash/main.js`
+- Create: `.obsidian/plugins/webinar-dash/styles.css`
+
+**Interfaces:**
+- Consumes: `derivePipeline`, `extractRawPath` from Task 2; `parseConfig` from Task 1
+- Produces:
+  - `async readPipeline(app, cfg) => { processed, unprocessed, orphaned }` as a method on the plugin class
+  - `renderLeftPane(container, pipeline, onIngest)` where `onIngest` is `(file, rowEl) => void`; Task 6 supplies the real handler, this task passes a no-op
+  - CSS class root `.webinar-dash` and the shared primitives `.wd-eyebrow`, `.wd-row`, `.wd-btn`, `.wd-mono`
+
+- [ ] **Step 1: Create `styles.css` with tesanti tokens**
+
+Values are copied verbatim from `colors_and_type.css`. Scoped to `.webinar-dash` so nothing leaks into the rest of Obsidian. Obsidian loads `styles.css` from the plugin folder automatically.
+
+```css
+.webinar-dash {
+  --wd-ink-000: #ffffff; --wd-ink-050: #f7f7f7; --wd-ink-100: #ececec;
+  --wd-ink-200: #d9d9d9; --wd-ink-400: #8a8a8a; --wd-ink-500: #5e5e5e;
+  --wd-ink-700: #262626; --wd-ink-900: #0a0a0a; --wd-ink-999: #000000;
+  --wd-red-500: #e1261c; --wd-red-600: #c31c14;
+
+  --wd-bg: var(--wd-ink-000);
+  --wd-bg-subtle: var(--wd-ink-050);
+  --wd-fg: var(--wd-ink-999);
+  --wd-fg-2: var(--wd-ink-700);
+  --wd-fg-3: var(--wd-ink-500);
+  --wd-fg-4: var(--wd-ink-400);
+  --wd-border: var(--wd-ink-200);
+  --wd-border-strong: var(--wd-ink-999);
+  --wd-border-subtle: var(--wd-ink-100);
+  --wd-accent: var(--wd-red-500);
+  --wd-accent-press: var(--wd-red-600);
+  --wd-accent-on: #ffffff;
+  --wd-ok: #0a8a3f;
+  --wd-warn: #c68a00;
+  --wd-danger: var(--wd-red-500);
+
+  --wd-mono: "JetBrains Mono", ui-monospace, SFMono-Regular, Menlo, Consolas, monospace;
+  --wd-sans: "Inter", ui-sans-serif, system-ui, -apple-system, "Helvetica Neue", Arial, sans-serif;
+
+  font-family: var(--wd-sans);
+  color: var(--wd-fg);
+  display: grid;
+  grid-template-columns: minmax(0, 38fr) minmax(0, 62fr);
+  gap: 24px;
+}
+
+.theme-dark .webinar-dash {
+  --wd-bg: var(--wd-ink-900);
+  --wd-bg-subtle: #161616;
+  --wd-fg: var(--wd-ink-000);
+  --wd-fg-2: var(--wd-ink-200);
+  --wd-fg-3: var(--wd-ink-400);
+  --wd-fg-4: var(--wd-ink-500);
+  --wd-border: var(--wd-ink-700);
+  --wd-border-strong: #b8b8b8;
+  --wd-border-subtle: #161616;
+  /* Accent tracks the installed theme at .obsidian/themes/tesanti/theme.css,
+     whose dark section reads "Black canvas, same signal red": --accent-h/-s/-l
+     are declared once at :root and never redeclared under .theme-dark, and only
+     the hover state lifts (#c31c14 light, #ff4d43 dark). Match that exactly. */
+  --wd-accent: var(--wd-red-500);
+  --wd-accent-press: #ff4d43;
+  --wd-accent-on: #ffffff;
+  /* Status tokens are separate from the brand accent by design-system rule, and
+     here they carry small text in a dense table. #e1261c on near-black is about
+     3.8:1, under AA for small text, so these lift where the accent does not. */
+  --wd-ok: #2fbf6a;
+  --wd-warn: #e0a516;
+  --wd-danger: #ff5c50;
+}
+
+@media (max-width: 820px) {
+  .webinar-dash { grid-template-columns: minmax(0, 1fr); }
+}
+
+.webinar-dash > * { min-width: 0; }
+
+.wd-pane { display: flex; flex-direction: column; gap: 20px; }
+
+.wd-eyebrow {
+  font-family: var(--wd-mono); font-size: 11px; line-height: 1;
+  letter-spacing: 0.12em; text-transform: uppercase;
+  color: var(--wd-fg-3); font-weight: 500;
+  display: flex; align-items: center; gap: 8px;
+}
+.wd-eyebrow::before {
+  content: ""; width: 5px; height: 5px; flex: none; background: var(--wd-accent);
+}
+
+.wd-block { display: flex; flex-direction: column; gap: 12px; }
+
+.wd-row {
+  display: flex; align-items: center; gap: 12px;
+  border: 1px solid var(--wd-border); border-radius: 4px;
+  padding: 12px 12px 12px 16px; background: var(--wd-bg);
+}
+.wd-row + .wd-row { margin-top: -1px; }
+.wd-row-main { flex: 1; min-width: 0; display: flex; flex-direction: column; gap: 2px; }
+.wd-row-name {
+  font-size: 14px; font-weight: 600;
+  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
+}
+.wd-mono {
+  font-family: var(--wd-mono); font-size: 11px; color: var(--wd-fg-3);
+  font-variant-numeric: tabular-nums;
+}
+
+.wd-btn {
+  display: inline-flex; align-items: center; gap: 6px; flex: none;
+  font-family: var(--wd-sans); font-size: 13px; font-weight: 600;
+  padding: 6px 12px; border-radius: 6px; cursor: pointer;
+  border: 1px solid var(--wd-accent); background: var(--wd-accent);
+  color: var(--wd-accent-on);
+  transition: background 120ms cubic-bezier(0.2, 0, 0, 1);
+}
+.wd-btn:hover { background: var(--wd-accent-press); border-color: var(--wd-accent-press); }
+.wd-btn:focus-visible { outline: 2px solid var(--wd-accent); outline-offset: 2px; }
+.wd-btn[disabled] {
+  opacity: 0.45; cursor: not-allowed;
+  background: transparent; color: var(--wd-fg-3); border-color: var(--wd-border);
+}
+.wd-btn svg { width: 14px; height: 14px; flex: none; }
+
+.wd-done { display: flex; flex-direction: column; }
+.wd-done-row {
+  display: flex; align-items: baseline; gap: 12px; padding: 5px 0;
+  border-bottom: 1px solid var(--wd-border-subtle); font-size: 13px;
+}
+.wd-done-row:last-child { border-bottom: 0; }
+.wd-done-name {
+  color: var(--wd-fg-2);
+  overflow: hidden; text-overflow: ellipsis; white-space: nowrap;
+}
+
+.wd-note { font-size: 12px; color: var(--wd-fg-3); }
+.wd-error {
+  font-size: 13px; color: var(--wd-danger);
+  border: 1px solid var(--wd-danger); border-radius: 4px; padding: 12px;
+}
+
+@media (prefers-reduced-motion: reduce) {
+  .webinar-dash * { transition-duration: 0.01ms !important; }
+}
+```
+
+- [ ] **Step 2: Add the Lucide icon helper and the left-pane renderer**
+
+Insert into `main.js` above the plugin class. Icons are inline stroke paths — the Lucide CDN is unavailable offline and the design system forbids emoji.
+
+```js
+// Source types CLAUDE.md documents for raw/sources. Anything else in that
+// folder is not a source and stays out of the queue.
+const RAW_EXTENSIONS = new Set(["md", "txt", "pdf"]);
+
+const ICONS = {
+  terminal: "M4 17l6-6-6-6M12 19h8",
+  check: "M20 6 9 17l-5-5",
+  minus: "M5 12h14",
+  x: "M18 6 6 18M6 6l12 12",
+  refresh: "M3 12a9 9 0 0 1 9-9 9 9 0 0 1 6.7 3L21 8M21 3v5h-5 M21 12a9 9 0 0 1-9 9 9 9 0 0 1-6.7-3L3 16M3 21v-5h5",
+};
+
+// Built with createElementNS rather than Obsidian's createSvg helper, whose
+// availability varies by version. This works on any Obsidian build.
+const SVG_NS = "http://www.w3.org/2000/svg";
+
+function addIcon(parent, name) {
+  const svg = document.createElementNS(SVG_NS, "svg");
+  svg.setAttribute("viewBox", "0 0 24 24");
+  svg.setAttribute("fill", "none");
+  svg.setAttribute("aria-hidden", "true");
+  const path = document.createElementNS(SVG_NS, "path");
+  path.setAttribute("d", ICONS[name]);
+  path.setAttribute("stroke", "currentColor");
+  path.setAttribute("stroke-width", "1.75");
+  path.setAttribute("stroke-linecap", "round");
+  path.setAttribute("stroke-linejoin", "round");
+  svg.appendChild(path);
+  parent.appendChild(svg);
+  return svg;
+}
+
+function formatBytes(n) {
+  return `${n.toLocaleString("en-US")} B`;
+}
+
+function renderLeftPane(container, pipeline, onIngest) {
+  const pane = container.createDiv({ cls: "wd-pane" });
+
+  const queue = pane.createDiv({ cls: "wd-block" });
+  queue.createDiv({
+    cls: "wd-eyebrow",
+    text: `Queue — ${pipeline.unprocessed.length} unprocessed`,
+  });
+  if (pipeline.unprocessed.length === 0) {
+    queue.createDiv({ cls: "wd-note", text: "Every raw source has a summary page." });
+  }
+  for (const file of pipeline.unprocessed) {
+    const row = queue.createDiv({ cls: "wd-row" });
+    const main = row.createDiv({ cls: "wd-row-main" });
+    main.createDiv({ cls: "wd-row-name", text: file.name });
+    main.createDiv({ cls: "wd-mono", text: formatBytes(file.size) });
+    const btn = row.createEl("button", { cls: "wd-btn" });
+    addIcon(btn, "terminal");
+    btn.createSpan({ text: "Ingest" });
+    btn.addEventListener("click", () => onIngest(file, row));
+  }
+
+  const done = pane.createDiv({ cls: "wd-block" });
+  done.createDiv({ cls: "wd-eyebrow", text: `Ingested — ${pipeline.processed.length}` });
+  const list = done.createDiv({ cls: "wd-done" });
+  for (const file of pipeline.processed) {
+    const row = list.createDiv({ cls: "wd-done-row" });
+    const date = file.page.name.slice(0, 10);
+    row.createSpan({ cls: "wd-mono", text: /^\d{4}-\d{2}-\d{2}$/.test(date) ? date : "—" });
+    row.createSpan({ cls: "wd-done-name", text: file.name.replace(/\.md$/, "") });
+  }
+
+  if (pipeline.orphaned.length > 0) {
+    const orphan = pane.createDiv({ cls: "wd-block" });
+    orphan.createDiv({
+      cls: "wd-eyebrow",
+      text: `Orphaned — ${pipeline.orphaned.length}`,
+    });
+    for (const page of pipeline.orphaned) {
+      const row = orphan.createDiv({ cls: "wd-done-row" });
+      row.createSpan({ cls: "wd-done-name", text: page.name });
+      row.createSpan({ cls: "wd-mono", text: page.rawPath || "no raw path" });
+    }
+  }
+
+  return pane;
+}
+```
+
+- [ ] **Step 3: Replace the plugin class body to read the vault and render**
+
+`vault.getFiles()` returns every markdown file; filter by path prefix. `vault.cachedRead` is the correct read for display purposes.
+
+```js
+class WebinarDashPlugin extends PluginBase {
+  async onload() {
+    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
+      const cfg = parseConfig(source);
+      const root = el.createDiv({ cls: "webinar-dash" });
+      try {
+        const pipeline = await this.readPipeline(cfg);
+        renderLeftPane(root, pipeline, () => {});
+      } catch (err) {
+        root.createDiv({ cls: "wd-error", text: `Dashboard failed: ${err.message}` });
+      }
+    });
+  }
+
+  async readPipeline(cfg) {
+    const all = this.app.vault.getFiles();
+    const rawDir = cfg.rawDir.replace(/\/+$/, "") + "/";
+    const wikiDir = cfg.wikiSourceDir.replace(/\/+$/, "") + "/";
+
+    // CLAUDE.md documents raw/sources as holding "markdown/text/pdf exports".
+    // Allow-listing only "md" would hide the others from the queue with no
+    // warning — the same silent-invisibility failure this dashboard exists to
+    // remove. Wiki source pages below stay markdown-only; those really are .md.
+    const rawFiles = all
+      .filter((f) => f.path.startsWith(rawDir) && RAW_EXTENSIONS.has(f.extension))
+      .map((f) => ({ path: f.path, name: f.name, size: f.stat.size }));
+
+    const pageFiles = all.filter((f) => f.path.startsWith(wikiDir) && f.extension === "md");
+    const sourcePages = [];
+    for (const f of pageFiles) {
+      const text = await this.app.vault.cachedRead(f);
+      sourcePages.push({ path: f.path, name: f.name, rawPath: extractRawPath(text) });
+    }
+
+    return derivePipeline({ rawFiles, sourcePages });
+  }
+}
+```
+
+- [ ] **Step 4: Verify in Obsidian**
+
+Reload the app (`Ctrl+R`) and open `dashboard.md` in reading view.
+
+Expected, against the vault's current state:
+- `Queue — 3 unprocessed` listing `Agentic Engineering, explained by a 10x developer.md` (15,020 B), `Webinar Plan - From Chat Box to Your Own OS.md` (17,177 B), and `Webinar script.md` (15,841 B), each with a red **Ingest** button that does nothing yet.
+- `Ingested — 9` listing the source summaries newest first, starting with `2026-07-24`.
+- No orphaned block.
+
+- [ ] **Step 5: Run the existing tests to confirm nothing regressed**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test
+```
+
+Bare `node --test` auto-discovers the test files. Do not pass `test/` as an
+argument — Node 24 resolves a bare directory path as a module and fails with
+`MODULE_NOT_FOUND` before running anything.
+
+Expected: PASS, 9 tests.
+
+- [ ] **Step 6: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add .obsidian/plugins/webinar-dash
+git commit -m "feat: render source pipeline in left pane"
+```
+
+---
+
+### Task 4: Coverage table parser
+
+**Files:**
+- Modify: `.obsidian/plugins/webinar-dash/main.js`
+- Test: `.obsidian/plugins/webinar-dash/test/coverage.test.js`
+
+**Interfaces:**
+- Consumes: `STATIONS` from Task 1
+- Produces:
+  - `parseCoverageTable(markdown) => { meta, rows, errors }`
+  - `meta`: `{ script: string | null, lastSynced: string | null }`
+  - `rows` items: `{ concept: string, status: "covered"|"partial"|"absent", stations: string[], pinned: boolean, line: number }`
+  - `errors` items: `{ line: number, text: string, reason: string }`
+  - `groupByStation(rows) => Array<{ station: string, rows: Row[] }>` ordered `All stations`, then the seven stations, then `No station`; empty groups omitted
+
+- [ ] **Step 1: Write the failing tests**
+
+Create `.obsidian/plugins/webinar-dash/test/coverage.test.js`.
+
+```js
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { parseCoverageTable, groupByStation } = require("../main.js").__test__;
+
+const DOC = [
+  "# Script coverage",
+  "",
+  "#coverage",
+  "",
+  "## Metadata",
+  "",
+  "- **Script:** `raw/sources/Webinar script.md`",
+  "- **Last synced:** 2026-07-28",
+  "",
+  "## Coverage",
+  "",
+  "| Concept | Status | Station | Pinned |",
+  "|---|---|---|---|",
+  "| [[harness]] | covered | Tools | |",
+  "| [[agentic-loops]] | partial | Process | |",
+  "| [[levels-of-ai-usage]] | partial | all | |",
+  "| [[connections-as-moat]] | absent | — | yes |",
+].join("\n");
+
+test("parseCoverageTable reads metadata", () => {
+  const { meta } = parseCoverageTable(DOC);
+  assert.equal(meta.script, "raw/sources/Webinar script.md");
+  assert.equal(meta.lastSynced, "2026-07-28");
+});
+
+test("parseCoverageTable reads every data row and skips the header", () => {
+  const { rows, errors } = parseCoverageTable(DOC);
+  assert.equal(errors.length, 0);
+  assert.equal(rows.length, 4);
+  assert.deepEqual(rows.map((r) => r.concept), [
+    "harness", "agentic-loops", "levels-of-ai-usage", "connections-as-moat",
+  ]);
+});
+
+test("parseCoverageTable normalises stations", () => {
+  const { rows } = parseCoverageTable(DOC);
+  assert.deepEqual(rows[0].stations, ["Tools"]);
+  assert.deepEqual(rows[2].stations, ["all"]);
+  assert.deepEqual(rows[3].stations, []);
+});
+
+test("parseCoverageTable reads the pinned flag", () => {
+  const { rows } = parseCoverageTable(DOC);
+  assert.equal(rows[0].pinned, false);
+  assert.equal(rows[3].pinned, true);
+});
+
+test("parseCoverageTable strips a wikilink alias", () => {
+  const doc = "| Concept | Status | Station |\n|---|---|---|\n| [[harness\\|The harness]] | covered | Tools |";
+  const { rows } = parseCoverageTable(doc);
+  assert.equal(rows[0].concept, "harness");
+});
+
+test("parseCoverageTable splits a multi-station cell", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered | Tools, Memory |";
+  const { rows } = parseCoverageTable(doc);
+  assert.deepEqual(rows[0].stations, ["Tools", "Memory"]);
+});
+
+test("parseCoverageTable rejects an invalid status instead of coercing it", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | maybe | Tools |";
+  const { rows, errors } = parseCoverageTable(doc);
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 1);
+  assert.match(errors[0].reason, /invalid status/);
+  assert.equal(errors[0].line, 3);
+});
+
+test("parseCoverageTable reports a row with too few columns", () => {
+  const doc = "| C | S | St |\n|---|---|---|\n| [[x]] | covered |";
+  const { rows, errors } = parseCoverageTable(doc);
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 1);
+  assert.match(errors[0].reason, /at least 3 columns/);
+});
+
+test("parseCoverageTable returns empty results for a document with no table", () => {
+  const { rows, errors } = parseCoverageTable("# Nothing\n\nJust prose.");
+  assert.equal(rows.length, 0);
+  assert.equal(errors.length, 0);
+});
+
+test("groupByStation orders all-stations first and no-station last", () => {
+  const { rows } = parseCoverageTable(DOC);
+  const groups = groupByStation(rows);
+  assert.deepEqual(groups.map((g) => g.station), [
+    "All stations", "Tools", "Process", "No station",
+  ]);
+  assert.equal(groups[1].rows[0].concept, "harness");
+});
+
+test("groupByStation places a multi-station row under its first station only", () => {
+  const rows = [{ concept: "x", status: "covered", stations: ["Memory", "Skills"], pinned: false, line: 1 }];
+  const groups = groupByStation(rows);
+  assert.equal(groups.length, 1);
+  assert.equal(groups[0].station, "Memory");
+});
+```
+
+- [ ] **Step 2: Run the tests to verify they fail**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test test/coverage.test.js
+```
+
+Expected: FAIL — `parseCoverageTable is not a function`.
+
+- [ ] **Step 3: Implement the parser**
+
+Insert into `main.js` below `derivePipeline`. The `inTable` flag flips on the `|---|` separator, which is what distinguishes the header row from data rows.
+
+```js
+const VALID_STATUS = new Set(["covered", "partial", "absent"]);
+const SEPARATOR_RE = /^\|?\s*:?-{2,}/;
+
+function parseCoverageTable(markdown) {
+  const text = String(markdown);
+  const meta = { script: null, lastSynced: null };
+
+  const scriptM = text.match(/^\s*-\s*\*\*Script:\*\*\s*`([^`]+)`/m);
+  if (scriptM) meta.script = scriptM[1].trim();
+  const syncM = text.match(/^\s*-\s*\*\*Last synced:\*\*\s*(\S+)/m);
+  if (syncM) meta.lastSynced = syncM[1].trim();
+
+  const rows = [];
+  const errors = [];
+  let inTable = false;
+
+  text.split(/\r?\n/).forEach((line, i) => {
+    const t = line.trim();
+    if (!t.startsWith("|")) {
+      inTable = false;
+      return;
+    }
+    if (SEPARATOR_RE.test(t)) {
+      inTable = true;
+      return;
+    }
+    if (!inTable) return;
+
+    const body = t.endsWith("|") ? t.slice(1, -1) : t.slice(1);
+    // Split on unescaped pipes only. Obsidian escapes the pipe of a piped
+    // wikilink inside a table cell as `\|`, and a plain split("|") tears
+    // `[[harness\|alias]]` into two cells, shifting every later column left —
+    // the status cell then reads "The harness]]" and the row is rejected.
+    const cells = body.split(/(?<!\\)\|/).map((c) => c.trim());
+
+    if (cells.length < 3) {
+      errors.push({ line: i + 1, text: t, reason: "expected at least 3 columns" });
+      return;
+    }
+
+    const status = cells[1].toLowerCase();
+    if (!VALID_STATUS.has(status)) {
+      errors.push({ line: i + 1, text: t, reason: `invalid status "${cells[1]}"` });
+      return;
+    }
+
+    // Capture up to the first ] | or backslash. Obsidian escapes the pipe in a
+    // piped wikilink inside a table cell, so the raw cell reads [[name\|alias]] —
+    // excluding the backslash is what keeps the trailing "\" out of the name.
+    const linkM = cells[0].match(/\[\[([^\]|\\]+)/);
+    const stationCell = cells[2];
+
+    rows.push({
+      concept: linkM ? linkM[1].trim() : cells[0],
+      status,
+      stations:
+        stationCell === "—" || stationCell === "-" || stationCell === ""
+          ? []
+          : stationCell.split(",").map((s) => s.trim()).filter(Boolean),
+      pinned: (cells[3] || "").toLowerCase() === "yes",
+      line: i + 1,
+    });
+  });
+
+  return { meta, rows, errors };
+}
+
+function groupByStation(rows) {
+  const order = ["All stations", ...STATIONS, "No station"];
+  const buckets = new Map(order.map((k) => [k, []]));
+
+  for (const row of rows) {
+    let key;
+    if (row.stations.includes("all")) key = "All stations";
+    else if (row.stations.length === 0) key = "No station";
+    else key = row.stations[0];
+    if (!buckets.has(key)) buckets.set(key, []);
+    buckets.get(key).push(row);
+  }
+
+  const known = order.filter((k) => buckets.get(k).length > 0);
+  const unknown = [...buckets.keys()].filter((k) => !order.includes(k) && buckets.get(k).length > 0);
+  return [...known, ...unknown].map((station) => ({ station, rows: buckets.get(station) }));
+}
+```
+
+The `\\|` in the alias test is an escaped pipe inside a JS string, which reaches the parser as a literal `|` — matching how Obsidian escapes piped wikilinks inside table cells.
+
+- [ ] **Step 4: Export them**
+
+```js
+module.exports.__test__ = {
+  parseConfig, extractRawPath, derivePipeline,
+  parseCoverageTable, groupByStation,
+  STATIONS, DEFAULTS,
+};
+```
+
+- [ ] **Step 5: Run the tests to verify they pass**
+
+```bash
+node --test test/coverage.test.js
+```
+
+Expected: PASS, 11 tests.
+
+- [ ] **Step 6: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add .obsidian/plugins/webinar-dash
+git commit -m "feat: parse the script coverage table"
+```
+
+---
+
+### Task 5: Coverage data file and CLAUDE.md contract
+
+This is the wiki-side task. It creates the file the renderer reads and installs the rule that keeps it current. No plugin code changes.
+
+**Files:**
+- Create: `wiki/script-coverage.md`
+- Modify: `CLAUDE.md` (Folder Convention, Tagging Rules, Standard Workflows, Operational Commands)
+- Modify: `index.md`
+- Modify: `log.md`
+
+**Interfaces:**
+- Consumes: the row format defined in Task 4
+- Produces: `wiki/script-coverage.md` conforming to that format, with one row per page in `wiki/concepts/`
+
+- [ ] **Step 1: Create `wiki/script-coverage.md`**
+
+Statuses below are the assessment made during design, read from `raw/sources/Webinar script.md` against each concept page. All 19 concept pages are present.
+
+```markdown
+# Script coverage
+
+#coverage
+
+## Metadata
+
+- **Script:** `raw/sources/Webinar script.md`
+- **Last synced:** 2026-07-28
+- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
+
+## Coverage
+
+| Concept | Status | Station | Pinned |
+|---|---|---|---|
+| [[harness]] | covered | Tools | |
+| [[evolution-of-agent-tooling]] | covered | Tools | |
+| [[skills-as-memory]] | covered | Skills | |
+| [[solve-first-then-skillify]] | covered | Skills | |
+| [[personal-ai-operating-system]] | covered | OS | |
+| [[context-as-scarce-resource]] | partial | Memory | |
+| [[agentic-loops]] | partial | Process | |
+| [[code-as-throwaway]] | partial | OS | |
+| [[levels-of-ai-usage]] | partial | all | |
+| [[integration-dead-ends]] | absent | — | |
+| [[leave-less-room-for-imagination]] | absent | — | |
+| [[product-ownership]] | absent | — | |
+| [[connections-as-moat]] | absent | — | |
+| [[network-from-a-standing-start]] | absent | — | |
+| [[seniority-and-the-junior-squeeze]] | absent | — | |
+| [[decoupling-identity-from-profession]] | absent | — | |
+| [[think-wider-not-bigger]] | absent | — | |
+| [[make-more-cheap-code]] | absent | — | |
+| [[enterprise-ai-reality]] | absent | — | |
+
+## Notes
+
+- Eight of the ten absent concepts are human-side or strategy-side, per the grouping in `index.md`. The script's spine is machine-side and it never reaches that material.
+- The remaining two absent concepts are machine-side: [[integration-dead-ends]] and [[leave-less-room-for-imagination]]. The script demonstrates the happy path, so it never reaches connector gating or spec ambiguity — the two ways the machine side fails in practice.
+- The `ReAct` station carries no wiki concept at all.
+- Set `Pinned` to `yes` on any row whose status is a deliberate decision. Sync will not touch it.
+
+## Related Pages
+
+- [[overview]]
+- `raw/sources/Webinar script.md`
+```
+
+- [ ] **Step 2: Add the file to the folder convention in `CLAUDE.md`**
+
+In the `## Folder Convention` code block, add below the `wiki/overview.md` line:
+
+```text
+  script-coverage.md  # machine-maintained: concept coverage vs the webinar script
+```
+
+- [ ] **Step 3: Add the tag row in `CLAUDE.md`**
+
+In the Folder → tag table under `## Tagging Rules`, add a row after the `wiki/overview.md` row:
+
+```text
+| `wiki/script-coverage.md` | `#coverage`    |
+```
+
+- [ ] **Step 4: Add Workflow D to `CLAUDE.md`**
+
+Append to `## Standard Workflows`, after Workflow C:
+
+```markdown
+### Workflow D: Sync Script Coverage
+
+Maintains `wiki/script-coverage.md` — one row per page in `wiki/concepts/`, judged
+against the script named in that file's `**Script:**` metadata field.
+
+1. Read `wiki/script-coverage.md` and note every row where `Pinned` is `yes`.
+2. Read the script and every page in `wiki/concepts/`.
+3. For each concept, decide `Status` and `Station`:
+   - `covered` — the script delivers the idea, whether or not it uses the page's name.
+   - `partial` — the script gestures at it but never lands it.
+   - `absent` — the script never reaches it.
+   - `Station` is one of the seven, a comma-separated list, `all`, or `—` when absent.
+4. **Never modify a row whose `Pinned` is `yes`** — not its status, not its station.
+   A pinned row is the user's judgment and outranks yours.
+5. Add rows for concept pages with no row. Remove rows whose concept page no longer exists.
+6. Update `**Last synced:**` to today.
+7. Update `index.md` and append a `sync` entry to `log.md`.
+
+Run this workflow:
+
+- at the end of any ingest that creates or modifies a page in `wiki/concepts/`
+- whenever the script file itself changes
+- on the explicit `sync script coverage` intent
+
+The coverage baseline is always the raw script. Ingesting the script into
+`wiki/sources/` does not change the baseline.
+```
+
+- [ ] **Step 5: Add the intent to `CLAUDE.md`**
+
+In `## Operational Commands (Natural Language)`, add to the supported intents list:
+
+```markdown
+- "sync script coverage"
+```
+
+- [ ] **Step 6: Update `index.md`**
+
+Add a section after `## Sources`, and correct the stale not-yet-ingested line.
+
+```markdown
+## Coverage
+
+- [[script-coverage]] — every concept vs `raw/sources/Webinar script.md`; 5 covered, 4 partial, 10 absent
+```
+
+Replace the existing `**Raw, not yet ingested:**` line with the accurate set — the two note files live in `raw/notes/`, not `raw/sources/`, and one raw source was missing entirely:
+
+```markdown
+**Raw, not yet ingested:** `raw/sources/Agentic Engineering, explained by a 10x developer.md` · `raw/sources/Webinar Plan - From Chat Box to Your Own OS.md` · `raw/sources/Webinar script.md`
+```
+
+- [ ] **Step 7: Append to `log.md`**
+
+```markdown
+## 2026-07-28 — sync (script coverage, initial)
+- Intent: sync script coverage
+- Input: first run of Workflow D, establishing `wiki/script-coverage.md`.
+- Pages created: [[script-coverage]] (coverage) — 19 rows, one per concept page.
+- Pages updated: `CLAUDE.md` (folder convention, `#coverage` tag row, Workflow D, new intent), `index.md` (new Coverage section; corrected the not-yet-ingested list — `Ideas for webinar.md` and `my theses.md` were listed under `raw/sources/` but live in `raw/notes/`, and `Agentic Engineering, explained by a 10x developer.md` was missing entirely).
+- Notes: Initial assessment is 5 covered, 4 partial, 10 absent. Eight of the ten absent are human-side or strategy-side per this file's own concept grouping; the other two — [[integration-dead-ends]] and [[leave-less-room-for-imagination]] — are machine-side, absent because the script demos the happy path and never reaches connector gating or spec ambiguity. The `ReAct` station carries no wiki concept. No concept pages were edited.
+- Next: decide whether the human-side cluster earns a station or is a deliberate cut; decide separately whether the two absent machine-side failure modes belong in the demo; pin the rows that are decided so future syncs leave them alone.
+```
+
+- [ ] **Step 8: Verify the file parses**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node -e "
+const fs = require('fs');
+const { parseCoverageTable } = require('./main.js').__test__;
+const md = fs.readFileSync('../../../wiki/script-coverage.md', 'utf8');
+const r = parseCoverageTable(md);
+console.log('rows:', r.rows.length, 'errors:', r.errors.length);
+console.log('meta:', r.meta);
+if (r.errors.length) { console.log(r.errors); process.exit(1); }
+if (r.rows.length !== 19) { console.log('expected 19 rows'); process.exit(1); }
+"
+```
+
+Expected: `rows: 19 errors: 0` and metadata showing the script path and `2026-07-28`.
+
+- [ ] **Step 9: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add wiki/script-coverage.md CLAUDE.md index.md log.md
+git commit -m "feat: add script coverage file and Workflow D sync contract"
+```
+
+---
+
+### Task 6: Right pane — coverage rendering
+
+**Files:**
+- Modify: `.obsidian/plugins/webinar-dash/main.js`
+- Modify: `.obsidian/plugins/webinar-dash/styles.css`
+
+**Interfaces:**
+- Consumes: `parseCoverageTable`, `groupByStation` from Task 4; `addIcon` from Task 3; `wiki/script-coverage.md` from Task 5
+- Produces:
+  - `reconcileConcepts(rows, conceptNames) => { rows, unsynced, stale }` where `unsynced` is concept names with no row and `stale` is rows whose concept page is gone
+  - `renderRightPane(container, parsed, reconciled)`
+
+- [ ] **Step 1: Add the reconcile helper and its tests**
+
+Append to `.obsidian/plugins/webinar-dash/test/coverage.test.js`:
+
+```js
+const { reconcileConcepts } = require("../main.js").__test__;
+
+test("reconcileConcepts finds concept pages with no row", () => {
+  const rows = [{ concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 }];
+  const out = reconcileConcepts(rows, ["harness", "brand-new-concept"]);
+  assert.deepEqual(out.unsynced, ["brand-new-concept"]);
+  assert.deepEqual(out.stale, []);
+});
+
+test("reconcileConcepts finds rows whose concept page is gone", () => {
+  const rows = [
+    { concept: "harness", status: "covered", stations: ["Tools"], pinned: false, line: 1 },
+    { concept: "deleted-idea", status: "absent", stations: [], pinned: false, line: 2 },
+  ];
+  const out = reconcileConcepts(rows, ["harness"]);
+  assert.deepEqual(out.stale.map((r) => r.concept), ["deleted-idea"]);
+  assert.deepEqual(out.unsynced, []);
+});
+```
+
+- [ ] **Step 2: Run the tests to verify they fail**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test test/coverage.test.js
+```
+
+Expected: FAIL — `reconcileConcepts is not a function`.
+
+- [ ] **Step 3: Implement `reconcileConcepts` and export it**
+
+Insert into `main.js` below `groupByStation`:
+
+```js
+function reconcileConcepts(rows, conceptNames) {
+  const named = new Set(conceptNames);
+  const rowed = new Set(rows.map((r) => r.concept));
+  return {
+    rows,
+    unsynced: conceptNames.filter((n) => !rowed.has(n)).sort(),
+    stale: rows.filter((r) => !named.has(r.concept)),
+  };
+}
+```
+
+Add `reconcileConcepts` to the `__test__` export object.
+
+- [ ] **Step 4: Run the tests to verify they pass**
+
+```bash
+node --test test/coverage.test.js
+```
+
+Expected: PASS, 13 tests.
+
+- [ ] **Step 5: Add the coverage styles**
+
+Append to `styles.css`:
+
+```css
+.wd-meter { display: flex; gap: 2px; height: 8px; width: 100%; }
+.wd-meter > i { display: block; height: 100%; }
+.wd-seg-covered { background: var(--wd-ok); }
+.wd-seg-partial { background: var(--wd-warn); }
+.wd-seg-absent { background: var(--wd-danger); }
+
+.wd-key { display: flex; flex-wrap: wrap; gap: 16px; font-size: 12px; color: var(--wd-fg-2); }
+.wd-key > span { display: inline-flex; align-items: center; gap: 6px; }
+.wd-key i { width: 8px; height: 8px; flex: none; }
+
+.wd-tbl { width: 100%; border-collapse: collapse; font-size: 13px; }
+.wd-tbl th {
+  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
+  text-transform: uppercase; color: var(--wd-fg-3); font-weight: 500;
+  text-align: left; padding: 0 12px 8px 0;
+  border-bottom: 1px solid var(--wd-border-strong);
+}
+.wd-tbl td {
+  padding: 6px 12px 6px 0; border-bottom: 1px solid var(--wd-border-subtle);
+  vertical-align: baseline;
+}
+.wd-grp td {
+  padding-top: 16px; border-bottom: 1px solid var(--wd-border-strong);
+  font-family: var(--wd-mono); font-size: 10px; letter-spacing: 0.1em;
+  text-transform: uppercase; color: var(--wd-fg); font-weight: 500;
+}
+.wd-stat { display: inline-flex; align-items: center; gap: 5px; font-size: 12px; font-weight: 600; }
+.wd-stat svg { width: 13px; height: 13px; flex: none; }
+.wd-covered { color: var(--wd-ok); }
+.wd-partial { color: var(--wd-warn); }
+.wd-absent { color: var(--wd-danger); }
+.wd-pin { font-family: var(--wd-mono); font-size: 10px; color: var(--wd-fg-4); }
+```
+
+- [ ] **Step 6: Add the right-pane renderer**
+
+Insert into `main.js` below `renderLeftPane`:
+
+```js
+const STATUS_ICON = { covered: "check", partial: "minus", absent: "x" };
+
+function renderRightPane(container, parsed, reconciled) {
+  const pane = container.createDiv({ cls: "wd-pane" });
+  pane.createDiv({ cls: "wd-eyebrow", text: "Coverage — by station" });
+
+  const counts = { covered: 0, partial: 0, absent: 0 };
+  for (const row of parsed.rows) counts[row.status] += 1;
+  const total = parsed.rows.length;
+
+  if (total > 0) {
+    const meter = pane.createDiv({ cls: "wd-meter" });
+    meter.setAttr(
+      "aria-label",
+      `Coverage: ${counts.covered} covered, ${counts.partial} partial, ${counts.absent} absent of ${total}`
+    );
+    for (const key of ["covered", "partial", "absent"]) {
+      if (counts[key] === 0) continue;
+      const seg = meter.createEl("i", { cls: `wd-seg-${key}` });
+      seg.style.flex = String(counts[key]);
+    }
+    const key = pane.createDiv({ cls: "wd-key" });
+    for (const name of ["covered", "partial", "absent"]) {
+      const span = key.createSpan();
+      span.createEl("i", { cls: `wd-seg-${name}` });
+      span.createSpan({ text: `${counts[name]} ${name}` });
+    }
+  }
+
+  if (parsed.errors.length > 0) {
+    const box = pane.createDiv({ cls: "wd-error" });
+    box.createDiv({ text: `${parsed.errors.length} unparseable row(s):` });
+    for (const e of parsed.errors) {
+      box.createDiv({ cls: "wd-mono", text: `line ${e.line} — ${e.reason}` });
+    }
+  }
+
+  const table = pane.createEl("table", { cls: "wd-tbl" });
+  const head = table.createEl("thead").createEl("tr");
+  for (const h of ["Concept", "Status", "Station", ""]) head.createEl("th", { text: h });
+  const body = table.createEl("tbody");
+
+  for (const group of groupByStation(parsed.rows)) {
+    const gr = body.createEl("tr", { cls: "wd-grp" });
+    gr.createEl("td", { attr: { colspan: "4" }, text: `${group.station} — ${group.rows.length}` });
+    for (const row of group.rows) {
+      const tr = body.createEl("tr");
+      // Obsidian's click handler resolves internal links via data-href, so both
+      // attributes are required for the link to open the concept page.
+      tr.createEl("td").createEl("a", {
+        cls: "internal-link",
+        text: row.concept,
+        attr: { href: row.concept, "data-href": row.concept },
+      });
+      const stat = tr.createEl("td").createSpan({ cls: `wd-stat wd-${row.status}` });
+      addIcon(stat, STATUS_ICON[row.status]);
+      stat.createSpan({ text: row.status });
+      tr.createEl("td", { cls: "wd-mono", text: row.stations.join(", ") || "—" });
+      tr.createEl("td", { cls: "wd-pin", text: row.pinned ? "pinned" : "" });
+    }
+  }
+
+  if (reconciled.unsynced.length > 0) {
+    const box = pane.createDiv({ cls: "wd-block" });
+    box.createDiv({ cls: "wd-eyebrow", text: `Unsynced — ${reconciled.unsynced.length}` });
+    box.createDiv({
+      cls: "wd-note",
+      text: `Concept pages with no row. Run "sync script coverage": ${reconciled.unsynced.join(", ")}`,
+    });
+  }
+
+  if (reconciled.stale.length > 0) {
+    const box = pane.createDiv({ cls: "wd-block" });
+    box.createDiv({ cls: "wd-eyebrow", text: `Stale — ${reconciled.stale.length}` });
+    box.createDiv({
+      cls: "wd-note",
+      text: `Rows whose concept page is gone: ${reconciled.stale.map((r) => r.concept).join(", ")}`,
+    });
+  }
+
+  return pane;
+}
+```
+
+- [ ] **Step 7: Wire it into the plugin class**
+
+Replace the code-block processor callback and add a reader method:
+
+```js
+    this.registerMarkdownCodeBlockProcessor("webinar-dash", async (source, el, ctx) => {
+      const cfg = parseConfig(source);
+      const root = el.createDiv({ cls: "webinar-dash" });
+      try {
+        const pipeline = await this.readPipeline(cfg);
+        renderLeftPane(root, pipeline, () => {});
+      } catch (err) {
+        root.createDiv({ cls: "wd-error", text: `Pipeline failed: ${err.message}` });
+      }
+      try {
+        const { parsed, reconciled } = await this.readCoverage(cfg);
+        renderRightPane(root, parsed, reconciled);
+      } catch (err) {
+        root.createDiv({ cls: "wd-error", text: `Coverage failed: ${err.message}` });
+      }
+    });
+```
+
+The two panes render in separate `try` blocks so a broken coverage file never blanks the source pipeline.
+
+```js
+  async readCoverage(cfg) {
+    const file = this.app.vault.getAbstractFileByPath(cfg.coverage);
+    if (!file) throw new Error(`no coverage file at ${cfg.coverage}`);
+    const parsed = parseCoverageTable(await this.app.vault.cachedRead(file));
+
+    const conceptDir = cfg.conceptDir.replace(/\/+$/, "") + "/";
+    const conceptNames = this.app.vault
+      .getFiles()
+      .filter((f) => f.path.startsWith(conceptDir) && f.extension === "md")
+      .map((f) => f.basename);
+
+    return { parsed, reconciled: reconcileConcepts(parsed.rows, conceptNames) };
+  }
+```
+
+- [ ] **Step 8: Verify in Obsidian**
+
+Reload (`Ctrl+R`) and open `dashboard.md`.
+
+Expected: a right pane with a meter reading `5 covered`, `4 partial`, `10 absent`; a table grouped `All stations — 1`, `Tools — 2`, `Memory — 1`, `Skills — 2`, `Process — 1`, `OS — 2`, `No station — 10`; no unsynced block; no stale block; concept names clickable through to their pages.
+
+- [ ] **Step 9: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add .obsidian/plugins/webinar-dash
+git commit -m "feat: render coverage meter and station-grouped table"
+```
+
+---
+
+### Task 7: Headless ingest
+
+**Files:**
+- Modify: `.obsidian/plugins/webinar-dash/main.js`
+- Modify: `.obsidian/plugins/webinar-dash/styles.css`
+- Test: `.obsidian/plugins/webinar-dash/test/safety.test.js`
+
+**Interfaces:**
+- Consumes: `renderLeftPane`'s `onIngest(file, rowEl)` callback from Task 3
+- Produces:
+  - `isSafeFilename(name) => boolean`
+  - `runIngest(file, rowEl)` as a plugin method
+
+- [ ] **Step 1: Write the failing tests**
+
+Create `.obsidian/plugins/webinar-dash/test/safety.test.js`. The accept cases are real filenames from this vault.
+
+```js
+"use strict";
+const test = require("node:test");
+const assert = require("node:assert/strict");
+const { isSafeFilename } = require("../main.js").__test__;
+
+test("isSafeFilename accepts every filename currently in the vault", () => {
+  const real = [
+    "Agentic Engineering, explained by a 10x developer.md",
+    "Webinar Plan - From Chat Box to Your Own OS.md",
+    "Webinar script.md",
+    "You're reading way too much code.md",
+    "ИИ глупый!.md",
+    "Скиллы на базе git — новая память AI-агентов.md",
+    "sebastian interview - conclusions and insights.md",
+    "In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md",
+  ];
+  for (const name of real) {
+    assert.equal(isSafeFilename(name), true, `should accept: ${name}`);
+  }
+});
+
+test("isSafeFilename rejects shell metacharacters", () => {
+  for (const bad of ['a".md', "a`b.md", "a$b.md", "a&b.md", "a|b.md", "a;b.md",
+                     "a<b.md", "a>b.md", "a%b.md", "a\nb.md", "a\rb.md"]) {
+    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
+  }
+});
+
+test("isSafeFilename rejects the cmd.exe escape character and control characters", () => {
+  // `^` escapes the next character in cmd.exe, so it can defuse the closing
+  // quote. NUL additionally makes spawn() throw synchronously.
+  for (const bad of ["a^b.md", "a\u0000b.md", "a\u001bb.md", "a\u007fb.md"]) {
+    assert.equal(isSafeFilename(bad), false, `should reject: ${JSON.stringify(bad)}`);
+  }
+});
+
+test("isSafeFilename rejects path traversal", () => {
+  assert.equal(isSafeFilename("../secrets.md"), false);
+  assert.equal(isSafeFilename("a/../../b.md"), false);
+});
+
+test("isSafeFilename rejects empty and non-string input", () => {
+  assert.equal(isSafeFilename(""), false);
+  assert.equal(isSafeFilename(null), false);
+  assert.equal(isSafeFilename(undefined), false);
+  assert.equal(isSafeFilename(42), false);
+});
+```
+
+- [ ] **Step 2: Run the tests to verify they fail**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test test/safety.test.js
+```
+
+Expected: FAIL — `isSafeFilename is not a function`.
+
+- [ ] **Step 3: Implement the validator and export it**
+
+Insert into `main.js` below `reconcileConcepts`:
+
+```js
+// `shell: true` is required on Windows to resolve `claude.cmd`, which puts the
+// filename into a shell string. Reject anything cmd.exe or a POSIX shell would
+// interpret, plus every control character.
+//
+// `^` is cmd.exe's escape character and belongs to the same class Node escapes
+// in its CVE-2024-27980 mitigation for exactly this spawn-through-.cmd shape.
+// Control characters are rejected because a NUL byte makes spawn() throw
+// synchronously, which would otherwise strand the row mid-run.
+//
+// Apostrophes, spaces, cyrillic, em dashes and `!` stay allowed — they are safe
+// inside double quotes and appear in real filenames in this vault. (`!` would
+// matter only under `setlocal enabledelayedexpansion`, which is not in play.)
+const UNSAFE_CHARS = /["`$&|;<>%^\u0000-\u001f\u007f]/;
+
+function isSafeFilename(name) {
+  if (typeof name !== "string" || name.length === 0) return false;
+  if (UNSAFE_CHARS.test(name)) return false;
+  if (name.includes("..")) return false;
+  return true;
+}
+```
+
+Add `isSafeFilename` to the `__test__` export object.
+
+- [ ] **Step 4: Run the tests to verify they pass**
+
+```bash
+node --test test/safety.test.js
+```
+
+Expected: PASS, 5 tests.
+
+- [ ] **Step 5: Add ingest-state styles**
+
+Append to `styles.css`:
+
+```css
+.wd-status { font-family: var(--wd-mono); font-size: 11px; flex: none; }
+.wd-status-running { color: var(--wd-warn); }
+.wd-status-done { color: var(--wd-ok); }
+.wd-status-failed { color: var(--wd-danger); }
+.wd-output {
+  font-family: var(--wd-mono); font-size: 11px; white-space: pre-wrap;
+  max-height: 220px; overflow: auto; margin-top: 8px;
+  border: 1px solid var(--wd-border); border-radius: 4px; padding: 8px;
+  color: var(--wd-fg-2);
+}
+```
+
+- [ ] **Step 6: Implement `runIngest` on the plugin class**
+
+`getBasePath()` exists on `FileSystemAdapter`; on mobile the adapter is a different class and `child_process` is unavailable, which is why the manifest sets `isDesktopOnly`.
+
+```js
+Add the in-flight registry as the first line of `onload()`, before the code-block
+processor is registered. It lives on the plugin instance so it survives the row
+re-renders that a DOM-scoped guard cannot:
+
+```js
+  async onload() {
+    this.running = new Set();
+    // ... existing registerMarkdownCodeBlockProcessor call follows unchanged
+```
+
+Then add both methods to the class:
+
+```js
+  vaultPath() {
+    const adapter = this.app.vault.adapter;
+    if (typeof adapter.getBasePath === "function") return adapter.getBasePath();
+    return null;
+  }
+
+  runIngest(file, rowEl) {
+    const Notice = OB ? OB.Notice : null;
+    const notify = (msg) => { if (Notice) new Notice(msg); };
+
+    // In-flight state is keyed on the file, not the row. Obsidian rebuilds the
+    // row on every re-render, so a row-scoped guard would let a re-render hand
+    // out a fresh row whose guard is unset — and a second click would then run
+    // a second unsupervised agent against the same file, concurrently writing
+    // the same wiki pages as the first.
+    if (this.running.has(file.path)) {
+      notify(`Already ingesting ${file.name}.`);
+      return;
+    }
+
+    if (!isSafeFilename(file.name)) {
+      rowEl.createDiv({
+        cls: "wd-output",
+        text: `Refused: "${file.name}" contains a character that is unsafe to pass to a shell.`,
+      });
+      return;
+    }
+
+    const base = this.vaultPath();
+    if (!base) {
+      notify("Ingest needs desktop Obsidian.");
+      return;
+    }
+
+    let spawn;
+    try {
+      ({ spawn } = require("child_process"));
+    } catch (_) {
+      notify("child_process unavailable — ingest needs desktop Obsidian.");
+      return;
+    }
+
+    // A retry reuses the same row. Clear the previous run's status and output
+    // so they are replaced rather than stacked on top of each other.
+    rowEl.querySelectorAll(".wd-status, .wd-output").forEach((el) => el.remove());
+
+    const button = rowEl.querySelector("button");
+    if (button) button.disabled = true;
+    this.running.add(file.path);
+
+    const status = rowEl.createSpan({ cls: "wd-status wd-status-running", text: "running 0s" });
+    const output = rowEl.createDiv({ cls: "wd-output", text: "" });
+    const started = Date.now();
+    const timer = window.setInterval(() => {
+      status.setText(`running ${Math.round((Date.now() - started) / 1000)}s`);
+    }, 1000);
+    this.registerInterval(timer);
+
+    // Bounded as it accumulates, not only when displayed, so a long or noisy
+    // run cannot grow this string without limit.
+    let buffered = "";
+    const append = (text) => {
+      buffered = (buffered + text).slice(-8000);
+      output.setText(buffered);
+      output.scrollTop = output.scrollHeight;
+    };
+
+    // Single exit path: every way this run can end clears the timer, releases
+    // the file, and re-enables the button.
+    const finish = (cls, label) => {
+      window.clearInterval(timer);
+      this.running.delete(file.path);
+      status.className = `wd-status ${cls}`;
+      status.setText(label);
+      if (button) button.disabled = false;
+    };
+
+    let child;
+    try {
+      child = spawn("claude", ["-p", `ingest "${file.name}"`], {
+        cwd: base,
+        shell: true,
+      });
+    } catch (err) {
+      // spawn() throws synchronously for some argument shapes. Without this the
+      // timer would run forever and the row would stay disabled until reload.
+      append(`\nCould not start claude: ${err.message}`);
+      finish("wd-status-failed", "failed");
+      return;
+    }
+
+    // Decode as UTF-8 across chunk boundaries. Raw Buffer chunks split wherever
+    // the OS buffer ends, and this vault's output is full of Cyrillic and em
+    // dashes that would otherwise decode as replacement characters.
+    child.stdout.setEncoding("utf8");
+    child.stderr.setEncoding("utf8");
+    child.stdout.on("data", append);
+    child.stderr.on("data", append);
+
+    child.on("error", (err) => {
+      append(`\nCould not start claude: ${err.message}\nIs it on PATH?`);
+      finish("wd-status-failed", "failed");
+    });
+
+    child.on("close", (code) => {
+      const secs = Math.round((Date.now() - started) / 1000);
+      if (code === 0) {
+        finish("wd-status-done", `done in ${secs}s`);
+        notify(`Ingested ${file.name}. Reopen the dashboard to refresh.`);
+      } else {
+        finish("wd-status-failed", `failed - exit ${code}`);
+      }
+    });
+  }
+```
+
+- [ ] **Step 7: Pass the real handler into the left pane**
+
+In the code-block processor, replace the no-op:
+
+```js
+        renderLeftPane(root, pipeline, (file, rowEl) => this.runIngest(file, rowEl));
+```
+
+- [ ] **Step 8: Verify the refusal path without spawning anything**
+
+Temporarily rename a raw file to include a shell metacharacter, reload, and click Ingest.
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/raw/sources"
+cp "Webinar script.md" 'bad&name.md'
+```
+
+Expected in Obsidian: the row shows `Refused: "bad&name.md" contains a character that is unsafe to pass to a shell.` and no process starts.
+
+Then remove it:
+
+```bash
+rm 'bad&name.md'
+```
+
+- [ ] **Step 9: Verify a real ingest**
+
+Reload Obsidian and click **Ingest** on `Agentic Engineering, explained by a 10x developer.md`.
+
+Expected: the button disables, the status ticks `running 1s`, `running 2s`, …, streamed output appears below the row, and on completion the status reads `done in Ns`. Reopen `dashboard.md`: the queue drops to 2 and the ingested list rises to 10.
+
+This writes to the vault unsupervised. Everything is committed, so `git diff HEAD` shows exactly what the agent changed and `git checkout -- .` reverts it.
+
+- [ ] **Step 10: Run the whole suite**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar/.obsidian/plugins/webinar-dash"
+node --test
+```
+
+Bare `node --test` auto-discovers the test files. Do not pass `test/` as an
+argument — Node 24 resolves a bare directory path as a module and fails with
+`MODULE_NOT_FOUND` before running anything.
+
+Expected: PASS, 27 tests across three files.
+
+- [ ] **Step 11: Commit**
+
+```bash
+cd "D:/Projects/Notes/Webinar/Webinar"
+git add .obsidian/plugins/webinar-dash
+git commit -m "feat: spawn headless claude ingest with filename validation"
+```
+
+---
+
+## Self-Review
+
+**Spec coverage.** Every section of the spec maps to a task:
+
+| Spec section | Task |
+|---|---|
+| Architecture, config block, no build step | 1 |
+| Source pipeline derivation, `**Raw path:**` join | 2 |
+| Left pane, tesanti styling, dark mode | 3 |
+| Coverage file format, parsing rules | 4 |
+| Coverage data file, four `CLAUDE.md` changes, `index.md`, `log.md` | 5 |
+| Two-pane layout, meter, station grouping, `unsynced` / `stale` | 6 |
+| Ingest spawn, filename rejection list, per-row states, preflight | 7 |
+| Failure modes table | 3 (orphaned), 6 (parse errors, unsynced, stale), 7 (no adapter, no `claude`) |
+
+The spec's "v1 does not write the coverage file from the plugin" is honored — no task adds a pin toggle.
+
+**Placeholder scan.** No `TBD`, no "add error handling", no "similar to Task N". Every code step carries runnable code.
+
+**Type consistency.** `rawFiles` items are `{path, name, size}` in Tasks 2 and 3. `sourcePages` items are `{path, name, rawPath}` in both. Coverage rows are `{concept, status, stations, pinned, line}` in Tasks 4, 5, and 6. `onIngest(file, rowEl)` is declared in Task 3 and implemented with the same signature in Task 7. `addIcon(parent, name)` is defined in Task 3 and reused in Task 6. `STATIONS` is defined in Task 1 and consumed by `groupByStation` in Task 4.
+
+**Test count.** Task 2 adds 9, Task 4 adds 11, Task 6 adds 2, Task 7 adds 5 — 27 total, matching Step 10 of Task 7.
diff --git a/docs/superpowers/specs/2026-07-28-webinar-dashboard-design.md b/docs/superpowers/specs/2026-07-28-webinar-dashboard-design.md
index 477195b..25fc39f 100644
--- a/docs/superpowers/specs/2026-07-28-webinar-dashboard-design.md
+++ b/docs/superpowers/specs/2026-07-28-webinar-dashboard-design.md
@@ -1,256 +1,256 @@
-# Webinar vault dashboard — design
-
-Date: 2026-07-28
-Status: approved, ready for implementation planning
-
-## Context
-
-The vault is an LLM-maintained wiki governed by `CLAUDE.md`. It currently holds 12 raw
-sources, 9 ingested source summaries, 19 concept pages, and one webinar script
-(`raw/sources/Webinar script.md`) that the concepts are supposed to feed.
-
-Two problems motivated this work:
-
-1. **The manual catalog drifts.** `index.md` lists `Ideas for webinar.md` and
-   `my theses.md` under `raw/sources/`, but both live in `raw/notes/`. It does not
-   mention `Agentic Engineering, explained by a 10x developer.md` at all, which sits
-   un-ingested in `raw/sources/`. Nothing detects this.
-2. **No view of script coverage.** There is no way to see which concept pages the
-   webinar script actually delivers. A manual read shows the script is entirely
-   machine-side: all human-side and strategy-side concepts are absent.
-
-## Goals
-
-- Show processed and unprocessed sources, with a one-click ingest on the unprocessed.
-- Show every concept and whether the webinar script mentions it, plus which script
-  station it lands in.
-- Keep coverage status in a separate markdown file, kept synchronized by a rule in
-  `CLAUDE.md`.
-
-## Non-goals
-
-- Replacing `index.md` or `log.md`. Both stay exactly as they are.
-- A general-purpose Obsidian dashboard framework. This is one vault-specific plugin.
-- Publishing the plugin to the community plugin registry.
-
-## Decisions taken
-
-| Decision | Choice | Rationale |
-|---|---|---|
-| Buttons | Own plugin, not Meta Bind | Needs are narrow and vault-specific; Meta Bind's expensive half (inline CM6 widgets, two-way frontmatter binding) is unused here |
-| Coverage status source | Claude judges; user can pin | Sync sets status automatically, but a row marked `Pinned: yes` is never overwritten |
-| Ingest mechanism | Headless `claude -p` via `child_process` | One click, fully automatic. Chosen over a queue file with the unsupervised-write trade-off understood |
-| Granularity | Status + script station | Turns the table into a pacing map, not just a checklist |
-| Layout | Two-pane (Option B) | Sources and coverage both first-class; coverage grouped by station recovers most of the station-board view |
-
-## Architecture
-
-```
-.obsidian/plugins/webinar-dash/
-  manifest.json
-  main.js          # plain CommonJS, no build step
-  styles.css       # tesanti tokens scoped to .webinar-dash
-
-dashboard.md               # vault root, beside index.md
-wiki/script-coverage.md    # the coverage table
-```
-
-`dashboard.md` holds only a config block; the plugin renders everything:
-
-````markdown
-# Webinar dashboard
-
-```webinar-dash
-script: raw/sources/Webinar script.md
-coverage: wiki/script-coverage.md
-```
-````
-
-Config keys are optional and fall back to those two defaults.
-
-### Where truth lives
-
-| Data | Source of truth | Mechanism |
-|---|---|---|
-| Which sources are processed | Filesystem, read live | Diff `raw/sources/*.md` against the `**Raw path:**` value in every `wiki/sources/*.md` |
-| Concept coverage status | `wiki/script-coverage.md` | Written by Claude on sync, read by the plugin |
-
-Mechanical facts come from the filesystem, judgment comes from the markdown file.
-This makes the `index.md` class of drift structurally impossible on the sources half:
-the dashboard cannot disagree with the filesystem because it derives from it.
-
-**The plugin never writes to the vault.** It reads files and spawns one subprocess.
-Nothing else. It also never reads `index.md` — the catalog is a human-facing artifact,
-and treating it as input would reintroduce exactly the drift this design removes.
-
-### Source pipeline derivation
-
-1. List `raw/sources/*.md`.
-2. For each `wiki/sources/*.md`, extract the backticked path from the line matching
-   `**Raw path:** \`<path>\``. Verified consistent across all 9 existing source pages.
-3. A raw file claimed by some source page is **processed**; unclaimed is **unprocessed**.
-4. A source page whose raw path no longer exists is reported as **orphaned**.
-
-`raw/notes/` is out of scope — those are notes, not sources.
-
-## Coverage file format
-
-```markdown
-# Script coverage
-
-#coverage
-
-## Metadata
-
-- **Script:** `raw/sources/Webinar script.md`
-- **Last synced:** 2026-07-28
-- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
-
-## Coverage
-
-| Concept | Status | Station | Pinned |
-|---|---|---|---|
-| [[harness]] | covered | Tools | |
-| [[agentic-loops]] | partial | Process | |
-| [[levels-of-ai-usage]] | partial | all | |
-| [[connections-as-moat]] | absent | — | yes |
-```
-
-Field rules:
-
-- **Concept** — an Obsidian wikilink to a page in `wiki/concepts/`. The plugin extracts
-  the page name from inside the brackets.
-- **Status** — exactly one of `covered`, `partial`, `absent`. Any other value renders as
-  `invalid` rather than being silently coerced.
-- **Station** — one of the seven station names, a comma-separated list of them, `all`,
-  or `—` for none.
-- **Pinned** — `yes`, or blank. Blank is the default.
-
-The seven stations are the `#` headings of the script that represent technology levels:
-Chat box, ReAct, Tools, Memory, Skills, Process, OS. The script's `Intro`,
-`Mail from boss`, and `Notes` headings are setup and are not stations.
-
-v1 does not write this file from the plugin. Pinning is a hand-edit of one cell — a pin
-toggle button would make the plugin a writer and risk clobbering concurrent user edits,
-which is not worth it for a one-word change.
-
-## Rendering — two-pane layout
-
-Left pane (38%):
-- **Queue** — unprocessed sources, each row showing filename, byte size, and an
-  **Ingest** button. Files missing from `index.md` need no special flag: they appear
-  here purely because no source page claims them, which is how
-  `Agentic Engineering, explained by a 10x developer.md` surfaces despite being absent
-  from the catalog.
-- **Ingested** — processed sources as compact rows: date from the source page filename
-  prefix, plus title.
-
-Right pane (62%):
-- Coverage meter — a stacked bar of covered / partial / absent with a 2px gap between
-  segments, plus a counted key.
-- Coverage table grouped by station, with `No station` last.
-
-Collapses to a single column below 820px so a narrow Obsidian pane stays usable.
-
-Styling follows the tesanti design system: black / white / red only, Space Grotesk
-display, Inter body, JetBrains Mono for eyebrows and data, radii at most 6px, 1px
-hairlines instead of shadows, Lucide stroke icons, no emoji. Status colors use the
-system's `--ok` / `--warn` / `--danger` tokens and always ship with a text label, never
-color alone. Dark mode is derived from the same ink ramp, with the red lifted to
-`#ff4a3d` so small text clears contrast on near-black.
-
-## Ingest mechanism
-
-```js
-const { spawn } = require("child_process");
-spawn("claude", ["-p", `ingest "${file}"`], { cwd: vaultPath, shell: true });
-```
-
-`shell: true` is required on Windows to resolve `claude.cmd`, which places the filename
-inside a shell string. Before spawning, the filename is rejected if it contains any of:
-`"`  `` ` ``  `$`  `&`  `|`  `;`  `<`  `>`  `%`  or a newline. Existing filenames include
-Cyrillic, spaces, and `!`, all of which pass. A rejected filename shows an error in its
-row and does not spawn.
-
-Vault path comes from `app.vault.adapter.getBasePath()` on `FileSystemAdapter`.
-
-Per-row states: `idle` → `running` with elapsed seconds → `done` or `failed · exit <n>`
-with captured stderr in an expandable block. On success the pipeline is re-derived and
-the row moves to the ingested list. A second click while running is ignored.
-
-A global **Sync coverage** button spawns `claude -p "sync script coverage"`.
-
-Preflight: if `claude` is not resolvable on PATH, all buttons render disabled with that
-reason stated.
-
-### Accepted risk
-
-Headless ingest writes source summaries, concept pages, `index.md`, and `log.md` without
-the user watching. This was chosen deliberately over a review checkpoint. Mitigations:
-one file per click rather than a batch, captured output retained per row, and visible
-per-row status. The writes land before the user reads them; this is understood and
-accepted.
-
-## CLAUDE.md changes
-
-1. **Folder convention** — add `wiki/script-coverage.md` to the tree with a note that it
-   is machine-maintained.
-2. **Tagging rules** — add the row `wiki/script-coverage.md` → `#coverage`. No existing
-   page type fits: it is generated tabular data, not prose analysis.
-3. **Workflow D: Sync Script Coverage** — re-read the script and every
-   `wiki/concepts/*.md`; set `Status` and `Station` for each; never modify a row whose
-   `Pinned` is `yes`; add rows for new concept pages; remove rows for deleted ones;
-   update `Last synced`; then update `index.md` and append to `log.md`.
-4. **Sync triggers** — Workflow D runs at the end of any ingest that creates or modifies
-   a concept page, whenever `Webinar script.md` changes, and on the explicit
-   `sync script coverage` intent, which is added to Operational Commands.
-
-The coverage baseline is always the raw script at `raw/sources/Webinar script.md`.
-Ingesting the script into `wiki/sources/` later does not change the baseline.
-
-## Failure modes
-
-| Condition | Behavior |
-|---|---|
-| Coverage file missing or table malformed | Sources pane renders normally; coverage pane shows a parse error with the offending line |
-| Concept page exists with no table row | Rendered as `unsynced`, so a stale sync is visible rather than silent |
-| Table row points at a nonexistent concept page | Rendered as `stale`, kept in place, not auto-removed |
-| Source page whose raw path is missing | Listed under `orphaned` in the left pane |
-| `child_process` unavailable (mobile) | Buttons render disabled with the reason |
-| `claude` not on PATH | Buttons render disabled with the reason |
-
-## Initial coverage assessment
-
-Read manually while designing; the first real sync will regenerate it. 19 concepts:
-5 covered, 4 partial, 10 absent.
-
-- **covered** — harness (Tools), skills-as-memory (Skills), solve-first-then-skillify
-  (Skills), personal-ai-operating-system (OS), evolution-of-agent-tooling (Tools)
-- **partial** — agentic-loops (Process), context-as-scarce-resource (Memory),
-  levels-of-ai-usage (all), code-as-throwaway (OS)
-- **absent** — connections-as-moat, product-ownership, seniority-and-the-junior-squeeze,
-  decoupling-identity-from-profession, network-from-a-standing-start,
-  think-wider-not-bigger, make-more-cheap-code, enterprise-ai-reality,
-  integration-dead-ends, leave-less-room-for-imagination
-
-Eight of the ten absent concepts are human-side or strategy-side per `index.md`'s grouping.
-The other two — `integration-dead-ends` and `leave-less-room-for-imagination` — are
-machine-side, and are absent because the script demonstrates the happy path and so never
-reaches connector gating or spec ambiguity, the two ways the machine side fails in
-practice. The `ReAct` station carries no wiki concept at all.
-
-## Out of scope for v1
-
-- Pin toggle button (hand-edit instead).
-- Alternate station-board view toggle. The data file is layout-independent, so this is a
-  render change if wanted later.
-- Coverage for entities, sources, or queries — concepts only.
-- Any view of `raw/notes/`.
-
-## Note on spec location
-
-This file introduces a `docs/` folder at the vault root, which is not part of the
-`CLAUDE.md` folder convention and will appear in Obsidian's file explorer. It can be
-moved or deleted without affecting the implementation.
-
-The vault is not a git repository, so this spec is not committed.
+# Webinar vault dashboard — design
+
+Date: 2026-07-28
+Status: approved, ready for implementation planning
+
+## Context
+
+The vault is an LLM-maintained wiki governed by `CLAUDE.md`. It currently holds 12 raw
+sources, 9 ingested source summaries, 19 concept pages, and one webinar script
+(`raw/sources/Webinar script.md`) that the concepts are supposed to feed.
+
+Two problems motivated this work:
+
+1. **The manual catalog drifts.** `index.md` lists `Ideas for webinar.md` and
+   `my theses.md` under `raw/sources/`, but both live in `raw/notes/`. It does not
+   mention `Agentic Engineering, explained by a 10x developer.md` at all, which sits
+   un-ingested in `raw/sources/`. Nothing detects this.
+2. **No view of script coverage.** There is no way to see which concept pages the
+   webinar script actually delivers. A manual read shows the script is entirely
+   machine-side: all human-side and strategy-side concepts are absent.
+
+## Goals
+
+- Show processed and unprocessed sources, with a one-click ingest on the unprocessed.
+- Show every concept and whether the webinar script mentions it, plus which script
+  station it lands in.
+- Keep coverage status in a separate markdown file, kept synchronized by a rule in
+  `CLAUDE.md`.
+
+## Non-goals
+
+- Replacing `index.md` or `log.md`. Both stay exactly as they are.
+- A general-purpose Obsidian dashboard framework. This is one vault-specific plugin.
+- Publishing the plugin to the community plugin registry.
+
+## Decisions taken
+
+| Decision | Choice | Rationale |
+|---|---|---|
+| Buttons | Own plugin, not Meta Bind | Needs are narrow and vault-specific; Meta Bind's expensive half (inline CM6 widgets, two-way frontmatter binding) is unused here |
+| Coverage status source | Claude judges; user can pin | Sync sets status automatically, but a row marked `Pinned: yes` is never overwritten |
+| Ingest mechanism | Headless `claude -p` via `child_process` | One click, fully automatic. Chosen over a queue file with the unsupervised-write trade-off understood |
+| Granularity | Status + script station | Turns the table into a pacing map, not just a checklist |
+| Layout | Two-pane (Option B) | Sources and coverage both first-class; coverage grouped by station recovers most of the station-board view |
+
+## Architecture
+
+```
+.obsidian/plugins/webinar-dash/
+  manifest.json
+  main.js          # plain CommonJS, no build step
+  styles.css       # tesanti tokens scoped to .webinar-dash
+
+dashboard.md               # vault root, beside index.md
+wiki/script-coverage.md    # the coverage table
+```
+
+`dashboard.md` holds only a config block; the plugin renders everything:
+
+````markdown
+# Webinar dashboard
+
+```webinar-dash
+script: raw/sources/Webinar script.md
+coverage: wiki/script-coverage.md
+```
+````
+
+Config keys are optional and fall back to those two defaults.
+
+### Where truth lives
+
+| Data | Source of truth | Mechanism |
+|---|---|---|
+| Which sources are processed | Filesystem, read live | Diff `raw/sources/*.md` against the `**Raw path:**` value in every `wiki/sources/*.md` |
+| Concept coverage status | `wiki/script-coverage.md` | Written by Claude on sync, read by the plugin |
+
+Mechanical facts come from the filesystem, judgment comes from the markdown file.
+This makes the `index.md` class of drift structurally impossible on the sources half:
+the dashboard cannot disagree with the filesystem because it derives from it.
+
+**The plugin never writes to the vault.** It reads files and spawns one subprocess.
+Nothing else. It also never reads `index.md` — the catalog is a human-facing artifact,
+and treating it as input would reintroduce exactly the drift this design removes.
+
+### Source pipeline derivation
+
+1. List `raw/sources/*.md`.
+2. For each `wiki/sources/*.md`, extract the backticked path from the line matching
+   `**Raw path:** \`<path>\``. Verified consistent across all 9 existing source pages.
+3. A raw file claimed by some source page is **processed**; unclaimed is **unprocessed**.
+4. A source page whose raw path no longer exists is reported as **orphaned**.
+
+`raw/notes/` is out of scope — those are notes, not sources.
+
+## Coverage file format
+
+```markdown
+# Script coverage
+
+#coverage
+
+## Metadata
+
+- **Script:** `raw/sources/Webinar script.md`
+- **Last synced:** 2026-07-28
+- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
+
+## Coverage
+
+| Concept | Status | Station | Pinned |
+|---|---|---|---|
+| [[harness]] | covered | Tools | |
+| [[agentic-loops]] | partial | Process | |
+| [[levels-of-ai-usage]] | partial | all | |
+| [[connections-as-moat]] | absent | — | yes |
+```
+
+Field rules:
+
+- **Concept** — an Obsidian wikilink to a page in `wiki/concepts/`. The plugin extracts
+  the page name from inside the brackets.
+- **Status** — exactly one of `covered`, `partial`, `absent`. Any other value renders as
+  `invalid` rather than being silently coerced.
+- **Station** — one of the seven station names, a comma-separated list of them, `all`,
+  or `—` for none.
+- **Pinned** — `yes`, or blank. Blank is the default.
+
+The seven stations are the `#` headings of the script that represent technology levels:
+Chat box, ReAct, Tools, Memory, Skills, Process, OS. The script's `Intro`,
+`Mail from boss`, and `Notes` headings are setup and are not stations.
+
+v1 does not write this file from the plugin. Pinning is a hand-edit of one cell — a pin
+toggle button would make the plugin a writer and risk clobbering concurrent user edits,
+which is not worth it for a one-word change.
+
+## Rendering — two-pane layout
+
+Left pane (38%):
+- **Queue** — unprocessed sources, each row showing filename, byte size, and an
+  **Ingest** button. Files missing from `index.md` need no special flag: they appear
+  here purely because no source page claims them, which is how
+  `Agentic Engineering, explained by a 10x developer.md` surfaces despite being absent
+  from the catalog.
+- **Ingested** — processed sources as compact rows: date from the source page filename
+  prefix, plus title.
+
+Right pane (62%):
+- Coverage meter — a stacked bar of covered / partial / absent with a 2px gap between
+  segments, plus a counted key.
+- Coverage table grouped by station, with `No station` last.
+
+Collapses to a single column below 820px so a narrow Obsidian pane stays usable.
+
+Styling follows the tesanti design system: black / white / red only, Space Grotesk
+display, Inter body, JetBrains Mono for eyebrows and data, radii at most 6px, 1px
+hairlines instead of shadows, Lucide stroke icons, no emoji. Status colors use the
+system's `--ok` / `--warn` / `--danger` tokens and always ship with a text label, never
+color alone. Dark mode is derived from the same ink ramp, with the red lifted to
+`#ff4a3d` so small text clears contrast on near-black.
+
+## Ingest mechanism
+
+```js
+const { spawn } = require("child_process");
+spawn("claude", ["-p", `ingest "${file}"`], { cwd: vaultPath, shell: true });
+```
+
+`shell: true` is required on Windows to resolve `claude.cmd`, which places the filename
+inside a shell string. Before spawning, the filename is rejected if it contains any of:
+`"`  `` ` ``  `$`  `&`  `|`  `;`  `<`  `>`  `%`  or a newline. Existing filenames include
+Cyrillic, spaces, and `!`, all of which pass. A rejected filename shows an error in its
+row and does not spawn.
+
+Vault path comes from `app.vault.adapter.getBasePath()` on `FileSystemAdapter`.
+
+Per-row states: `idle` → `running` with elapsed seconds → `done` or `failed · exit <n>`
+with captured stderr in an expandable block. On success the pipeline is re-derived and
+the row moves to the ingested list. A second click while running is ignored.
+
+A global **Sync coverage** button spawns `claude -p "sync script coverage"`.
+
+Preflight: if `claude` is not resolvable on PATH, all buttons render disabled with that
+reason stated.
+
+### Accepted risk
+
+Headless ingest writes source summaries, concept pages, `index.md`, and `log.md` without
+the user watching. This was chosen deliberately over a review checkpoint. Mitigations:
+one file per click rather than a batch, captured output retained per row, and visible
+per-row status. The writes land before the user reads them; this is understood and
+accepted.
+
+## CLAUDE.md changes
+
+1. **Folder convention** — add `wiki/script-coverage.md` to the tree with a note that it
+   is machine-maintained.
+2. **Tagging rules** — add the row `wiki/script-coverage.md` → `#coverage`. No existing
+   page type fits: it is generated tabular data, not prose analysis.
+3. **Workflow D: Sync Script Coverage** — re-read the script and every
+   `wiki/concepts/*.md`; set `Status` and `Station` for each; never modify a row whose
+   `Pinned` is `yes`; add rows for new concept pages; remove rows for deleted ones;
+   update `Last synced`; then update `index.md` and append to `log.md`.
+4. **Sync triggers** — Workflow D runs at the end of any ingest that creates or modifies
+   a concept page, whenever `Webinar script.md` changes, and on the explicit
+   `sync script coverage` intent, which is added to Operational Commands.
+
+The coverage baseline is always the raw script at `raw/sources/Webinar script.md`.
+Ingesting the script into `wiki/sources/` later does not change the baseline.
+
+## Failure modes
+
+| Condition | Behavior |
+|---|---|
+| Coverage file missing or table malformed | Sources pane renders normally; coverage pane shows a parse error with the offending line |
+| Concept page exists with no table row | Rendered as `unsynced`, so a stale sync is visible rather than silent |
+| Table row points at a nonexistent concept page | Rendered as `stale`, kept in place, not auto-removed |
+| Source page whose raw path is missing | Listed under `orphaned` in the left pane |
+| `child_process` unavailable (mobile) | Buttons render disabled with the reason |
+| `claude` not on PATH | Buttons render disabled with the reason |
+
+## Initial coverage assessment
+
+Read manually while designing; the first real sync will regenerate it. 19 concepts:
+5 covered, 4 partial, 10 absent.
+
+- **covered** — harness (Tools), skills-as-memory (Skills), solve-first-then-skillify
+  (Skills), personal-ai-operating-system (OS), evolution-of-agent-tooling (Tools)
+- **partial** — agentic-loops (Process), context-as-scarce-resource (Memory),
+  levels-of-ai-usage (all), code-as-throwaway (OS)
+- **absent** — connections-as-moat, product-ownership, seniority-and-the-junior-squeeze,
+  decoupling-identity-from-profession, network-from-a-standing-start,
+  think-wider-not-bigger, make-more-cheap-code, enterprise-ai-reality,
+  integration-dead-ends, leave-less-room-for-imagination
+
+Eight of the ten absent concepts are human-side or strategy-side per `index.md`'s grouping.
+The other two — `integration-dead-ends` and `leave-less-room-for-imagination` — are
+machine-side, and are absent because the script demonstrates the happy path and so never
+reaches connector gating or spec ambiguity, the two ways the machine side fails in
+practice. The `ReAct` station carries no wiki concept at all.
+
+## Out of scope for v1
+
+- Pin toggle button (hand-edit instead).
+- Alternate station-board view toggle. The data file is layout-independent, so this is a
+  render change if wanted later.
+- Coverage for entities, sources, or queries — concepts only.
+- Any view of `raw/notes/`.
+
+## Note on spec location
+
+This file introduces a `docs/` folder at the vault root, which is not part of the
+`CLAUDE.md` folder convention and will appear in Obsidian's file explorer. It can be
+moved or deleted without affecting the implementation.
+
+The vault is not a git repository, so this spec is not committed.
diff --git a/index.md b/index.md
index 7e69d5c..e078de1 100644
--- a/index.md
+++ b/index.md
@@ -1,89 +1,89 @@
-# Wiki Index
-
-Content catalog for this vault. Updated after every ingest / query / lint operation that changes the wiki.
-
-See [[overview]] for the top-level synthesis and navigation.
-
-Every wiki page carries a page-type tag under its H1 (`#source`, `#entity`, `#concept`, `#timeline`, `#comparison`, `#query`, `#lint-report`, `#overview`) — see "Tagging Rules" in `CLAUDE.md`.
-
-## Sources
-
-- [[2026-07-14-everything-we-knew-about-software-has-changed]] — Theo Browne (AIE): model eras, think wider, code as throwaway _(raw: Everything we knew about software has changed.md)_
-- [[2026-07-14-gap-between-ai-users-irreversible]] — Allie Miller: personal AI OS, foundation docs, skills, proactive workflows _(raw: In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md)_
-- [[2026-07-14-sebastian-eugene-interview]] — Sebastian + Eugene: harness, team collapse, connections, enterprise reality _(raw: sebastian interview - conclusions and insights.md)_
-- [[2026-07-14-skills-based-on-git]] — Konstantin (Sber): git skills as agent memory, harness, agent loops _(raw: Скиллы на базе git — новая память AI-агентов.md)_
-- [[2026-07-14-nina-interview]] — Eugene + Nina (HR): transcript > summary, friction not resistance, skills as handoff _(raw: Nina interview.md)_
-- [[2026-07-14-yulia-interview]] — Eugene + Yulia (HR lead): levels of AI usage, candidate knowledge base, solve-first _(raw: Yulia interview.md)_
-- [[2026-07-21-larysa-interview]] — Eugene + Larysa (BA/PM): memory loss, integration dead-ends, less room for imagination _(raw: Larysa interview.md)_
-- [[2026-07-22-ai-is-stupid]] — YouTube short (author unknown): "stupid AI" = model minus context minus harness; Nobel-vs-employee analogy _(raw: ИИ глупый!.md)_
-- [[2026-07-24-youre-reading-way-too-much-code]] — Theo Browne (video): make more cheap code, four tiers of code, 100:1 slop-to-ship verification _(raw: You're reading way too much code.md)_
-
-**Raw, not yet ingested:** `raw/sources/Agentic Engineering, explained by a 10x developer.md` · `raw/sources/Webinar Plan - From Chat Box to Your Own OS.md` · `raw/sources/Webinar script.md`
-
-## Coverage
-
-- [[script-coverage]] — every concept vs `raw/sources/Webinar script.md`; 5 covered, 4 partial, 10 absent
-
-## Entities
-
-### People
-- [[theo-browne]] — developer/educator (t3.gg); "think wider", "make more cheap code"
-- [[allie-miller]] — ex-Amazon AI leader; personal AI OS
-- [[sebastian]] — founder of Virtido; enterprise/connections lens
-- [[eugene]] — developer, harness-builder, webinar author; interviewer (likely vault owner, tentative)
-- [[konstantin]] — Sber/GigaChat R&D; skills-as-memory
-- [[nina]] — HR recruiter at Virtido; webinar-audience proxy, use-case supplier
-- [[yulia]] — HR/recruiting lead; webinar organizer (name/affiliation tentative)
-- [[larysa]] — technical BA/PM, ex-mobile dev; advanced user blocked by memory + integrations
-
-### Tools / Orgs
-- [[claude-code]] — reference harness (all sources)
-- [[hermes]] — skills-first, self-curating harness
-- [[virtido]] — Sebastian's outsourcing company; its HR team is the webinar audience
-- [[inspectron]] — Eugene's employer (Edge Compute / IoT)
-
-## Concepts
-
-**Machine side**
-- [[harness]] — universal agent = LLM + small toolset + loop
-- [[skills-as-memory]] — git skills (SKILL.md + tools + data) as agent memory
-- [[evolution-of-agent-tooling]] — tools → MCP → skills
-- [[agentic-loops]] — inner (ReAct) / outer (Ralph) / meta
-- [[context-as-scarce-resource]] — the binding constraint; the "smart zone"
-- [[personal-ai-operating-system]] — Allie's context docs + skills + proactive workflows
-- [[levels-of-ai-usage]] — Eugene's ladder: chatbot → … → CLAUDE.md + skills (the non-programmer ceiling)
-- [[solve-first-then-skillify]] — solve the task once, then freeze it into a skill
-- [[integration-dead-ends]] — agent starts work against connectors the user's account doesn't have
-- [[leave-less-room-for-imagination]] — every gap in a spec gets filled, invisibly; tighten it
-
-**Human side**
-- [[product-ownership]] — own outcomes, frame problems not tickets
-- [[connections-as-moat]] — in-person relationships as the last non-commoditized asset
-- [[network-from-a-standing-start]] — tentative from-zero networking protocol (v0, to be validated)
-- [[seniority-and-the-junior-squeeze]] — judgment as risk-reduction
-- [[decoupling-identity-from-profession]] — separate who you are from what you do
-
-**Strategy side**
-- [[think-wider-not-bigger]] — breadth over depth; match ambition to the model
-- [[code-as-throwaway]] — cost of code → zero
-- [[make-more-cheap-code]] — Theo: four tiers of code; generate never-shipped slop to verify/explore; there's always another layer
-- [[enterprise-ai-reality]] — compliance lock-down; the company-managed-harness market
-
-## Timelines
-
-- [[ai-agent-evolution]] — agent/tooling timeline + Theo's model eras
-
-## Comparisons
-
-- [[theo-konstantin-allie]] — Theo vs Konstantin vs Allie: three lenses (strategy / engineering / personal-OS) on the same shift; shared markdown-as-unit and system-over-model, differing altitude
-
-## Queries
-
-- [[2026-07-14-best-first-skill-for-beginner]] — best first skill for a Claude beginner: skill-creator (meta) + tone-of-voice/anti-AI-language; foundation docs first
-- [[2026-07-14-network-from-standing-start]] — network-from-zero: tentative protocol + Sebastian round-2 interview instrument (10 questions)
-- [[2026-07-22-webinar-theses]] — 14 candidate theses for the webinar, grouped spine / stakes / obstacles / method / tensions
-- [[2026-07-24-non-engineer-throwaway-verification]] — non-engineer analog of throwaway verification code: generated checks not content; checker skills; drift as diagnostic
-
-## Lint Reports
-
-_None yet._
+# Wiki Index
+
+Content catalog for this vault. Updated after every ingest / query / lint operation that changes the wiki.
+
+See [[overview]] for the top-level synthesis and navigation.
+
+Every wiki page carries a page-type tag under its H1 (`#source`, `#entity`, `#concept`, `#timeline`, `#comparison`, `#query`, `#lint-report`, `#overview`) — see "Tagging Rules" in `CLAUDE.md`.
+
+## Sources
+
+- [[2026-07-14-everything-we-knew-about-software-has-changed]] — Theo Browne (AIE): model eras, think wider, code as throwaway _(raw: Everything we knew about software has changed.md)_
+- [[2026-07-14-gap-between-ai-users-irreversible]] — Allie Miller: personal AI OS, foundation docs, skills, proactive workflows _(raw: In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md)_
+- [[2026-07-14-sebastian-eugene-interview]] — Sebastian + Eugene: harness, team collapse, connections, enterprise reality _(raw: sebastian interview - conclusions and insights.md)_
+- [[2026-07-14-skills-based-on-git]] — Konstantin (Sber): git skills as agent memory, harness, agent loops _(raw: Скиллы на базе git — новая память AI-агентов.md)_
+- [[2026-07-14-nina-interview]] — Eugene + Nina (HR): transcript > summary, friction not resistance, skills as handoff _(raw: Nina interview.md)_
+- [[2026-07-14-yulia-interview]] — Eugene + Yulia (HR lead): levels of AI usage, candidate knowledge base, solve-first _(raw: Yulia interview.md)_
+- [[2026-07-21-larysa-interview]] — Eugene + Larysa (BA/PM): memory loss, integration dead-ends, less room for imagination _(raw: Larysa interview.md)_
+- [[2026-07-22-ai-is-stupid]] — YouTube short (author unknown): "stupid AI" = model minus context minus harness; Nobel-vs-employee analogy _(raw: ИИ глупый!.md)_
+- [[2026-07-24-youre-reading-way-too-much-code]] — Theo Browne (video): make more cheap code, four tiers of code, 100:1 slop-to-ship verification _(raw: You're reading way too much code.md)_
+
+**Raw, not yet ingested:** `raw/sources/Agentic Engineering, explained by a 10x developer.md` · `raw/sources/Webinar Plan - From Chat Box to Your Own OS.md` · `raw/sources/Webinar script.md`
+
+## Coverage
+
+- [[script-coverage]] — every concept vs `raw/sources/Webinar script.md`; 5 covered, 4 partial, 10 absent
+
+## Entities
+
+### People
+- [[theo-browne]] — developer/educator (t3.gg); "think wider", "make more cheap code"
+- [[allie-miller]] — ex-Amazon AI leader; personal AI OS
+- [[sebastian]] — founder of Virtido; enterprise/connections lens
+- [[eugene]] — developer, harness-builder, webinar author; interviewer (likely vault owner, tentative)
+- [[konstantin]] — Sber/GigaChat R&D; skills-as-memory
+- [[nina]] — HR recruiter at Virtido; webinar-audience proxy, use-case supplier
+- [[yulia]] — HR/recruiting lead; webinar organizer (name/affiliation tentative)
+- [[larysa]] — technical BA/PM, ex-mobile dev; advanced user blocked by memory + integrations
+
+### Tools / Orgs
+- [[claude-code]] — reference harness (all sources)
+- [[hermes]] — skills-first, self-curating harness
+- [[virtido]] — Sebastian's outsourcing company; its HR team is the webinar audience
+- [[inspectron]] — Eugene's employer (Edge Compute / IoT)
+
+## Concepts
+
+**Machine side**
+- [[harness]] — universal agent = LLM + small toolset + loop
+- [[skills-as-memory]] — git skills (SKILL.md + tools + data) as agent memory
+- [[evolution-of-agent-tooling]] — tools → MCP → skills
+- [[agentic-loops]] — inner (ReAct) / outer (Ralph) / meta
+- [[context-as-scarce-resource]] — the binding constraint; the "smart zone"
+- [[personal-ai-operating-system]] — Allie's context docs + skills + proactive workflows
+- [[levels-of-ai-usage]] — Eugene's ladder: chatbot → … → CLAUDE.md + skills (the non-programmer ceiling)
+- [[solve-first-then-skillify]] — solve the task once, then freeze it into a skill
+- [[integration-dead-ends]] — agent starts work against connectors the user's account doesn't have
+- [[leave-less-room-for-imagination]] — every gap in a spec gets filled, invisibly; tighten it
+
+**Human side**
+- [[product-ownership]] — own outcomes, frame problems not tickets
+- [[connections-as-moat]] — in-person relationships as the last non-commoditized asset
+- [[network-from-a-standing-start]] — tentative from-zero networking protocol (v0, to be validated)
+- [[seniority-and-the-junior-squeeze]] — judgment as risk-reduction
+- [[decoupling-identity-from-profession]] — separate who you are from what you do
+
+**Strategy side**
+- [[think-wider-not-bigger]] — breadth over depth; match ambition to the model
+- [[code-as-throwaway]] — cost of code → zero
+- [[make-more-cheap-code]] — Theo: four tiers of code; generate never-shipped slop to verify/explore; there's always another layer
+- [[enterprise-ai-reality]] — compliance lock-down; the company-managed-harness market
+
+## Timelines
+
+- [[ai-agent-evolution]] — agent/tooling timeline + Theo's model eras
+
+## Comparisons
+
+- [[theo-konstantin-allie]] — Theo vs Konstantin vs Allie: three lenses (strategy / engineering / personal-OS) on the same shift; shared markdown-as-unit and system-over-model, differing altitude
+
+## Queries
+
+- [[2026-07-14-best-first-skill-for-beginner]] — best first skill for a Claude beginner: skill-creator (meta) + tone-of-voice/anti-AI-language; foundation docs first
+- [[2026-07-14-network-from-standing-start]] — network-from-zero: tentative protocol + Sebastian round-2 interview instrument (10 questions)
+- [[2026-07-22-webinar-theses]] — 14 candidate theses for the webinar, grouped spine / stakes / obstacles / method / tensions
+- [[2026-07-24-non-engineer-throwaway-verification]] — non-engineer analog of throwaway verification code: generated checks not content; checker skills; drift as diagnostic
+
+## Lint Reports
+
+_None yet._
diff --git a/log.md b/log.md
index 9313e90..171c562 100644
--- a/log.md
+++ b/log.md
@@ -1,157 +1,157 @@
-# Operation Log
-
-Append-only chronological record of wiki operations. Newest entries at the bottom.
-
-Entry format:
-
-```
-## YYYY-MM-DD — <operation>
-- Intent: ingest | query | lint | maintenance | sync
-- Input: <source path / question / scope>
-- Pages changed: [[...]], [[...]]
-- Notes: <what changed, uncertainty, next steps>
-```
-
----
-
-## 2026-07-14 — maintenance
-- Intent: maintenance
-- Input: Initialize vault folder structure.
-- Pages changed: created `index.md`, `log.md`, `wiki/overview.md`; created `raw/{sources,assets}` and `wiki/{sources,entities,concepts,timelines,comparisons,queries,lint-reports}`.
-- Notes: Empty vault scaffolded per the LLM Wiki Schema in `CLAUDE.md`. No sources ingested yet. Ready for first "ingest <source>".
-
-## 2026-07-14 — ingest (batch, 4 sources)
-- Intent: ingest
-- Input: The 4 talk/interview docs (user-selected scope; "Ideas for webinar" + "HR Contacts" left as raw reference).
-  - `raw/sources/Everything we knew about software has changed.md`
-  - `raw/sources/In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md`
-  - `raw/sources/sebastian interview - conclusions and insights.md`
-  - `raw/sources/Скиллы на базе git — новая память AI-агентов.md`
-- Pages created — sources (4): [[2026-07-14-everything-we-knew-about-software-has-changed]], [[2026-07-14-gap-between-ai-users-irreversible]], [[2026-07-14-sebastian-eugene-interview]], [[2026-07-14-skills-based-on-git]].
-- Pages created — entities (8): [[theo-browne]], [[allie-miller]], [[sebastian]], [[eugene]], [[konstantin]], [[claude-code]], [[hermes]], [[virtido]].
-- Pages created — concepts (13): [[harness]], [[skills-as-memory]], [[evolution-of-agent-tooling]], [[agentic-loops]], [[context-as-scarce-resource]], [[personal-ai-operating-system]], [[product-ownership]], [[connections-as-moat]], [[seniority-and-the-junior-squeeze]], [[decoupling-identity-from-profession]], [[think-wider-not-bigger]], [[code-as-throwaway]], [[enterprise-ai-reality]].
-- Pages created — timelines (1): [[ai-agent-evolution]].
-- Pages updated: [[overview]] (added the cross-source through-line + agree/diverge), `index.md` (full catalog).
-- Notes: Strong cross-source convergence (skills-as-memory: Konstantin ↔ Allie; connections: Sebastian ↔ Allie; code-as-throwaway: Theo ↔ Sebastian). Recorded contradictions (Eugene vs Sebastian on BYO- vs company-managed harness, online vs in-person networking, OSS motivation). Tentative claims flagged: Theo's model-era names, Eugene = vault owner, Hermes/Ralph heuristics, Allie's self-learning-models prediction.
-- Next: consider ingesting "Ideas for webinar" (connects corpus → webinar plan) and "HR Contacts" (the HR-search demo); consider a `wiki/comparisons/` page for the BYO-vs-managed-harness debate; run "lint wiki" to check link health.
-
-## 2026-07-14 — maintenance (tagging convention)
-- Intent: maintenance
-- Input: user request — add per-type hashtags to the wiki schema and existing pages.
-- Pages changed: `CLAUDE.md` (new "Tagging Rules" section, rule 9, templates + lint checklist updated); all 26 wiki pages received a page-type tag line under their H1 — [[overview]] `#overview`, 4 source pages `#source`, 8 entity pages `#entity`, 13 concept pages `#concept`, [[ai-agent-evolution]] `#timeline`; `index.md` (noted the convention).
-- Notes: Tag mapping is folder-driven (`#source`, `#entity`, `#concept`, `#timeline`, `#comparison`, `#query`, `#lint-report`, `#overview`). Exactly one type tag per page, first on the tag line; topical tags may follow. Lint now checks for missing/mismatched type tags.
-
-## 2026-07-14 -- query
-- Intent: query
-- Input: "what would be the best claude skill to give to a claude beginner user?"
-- Pages changed: created [[2026-07-14-best-first-skill-for-beginner]]; updated [[skills-as-memory]] (Next Questions item answered + linked), `index.md` (Queries section).
-- Notes: Answer synthesized from Allie Miller (skill-creator, tone-of-voice/anti-AI-language starter set, foundation-docs-first) and Konstantin (>5-tool-calls auto-creation heuristic). Follow-up: contents of a good tone-of-voice SKILL.md; foundation-doc templates.
-
-## 2026-07-14 — maintenance (authorized raw edit)
-- Intent: maintenance
-- Input: user explicitly authorized editing `raw/sources/Webinar script.md` this session (exception to rule 1).
-- Pages changed: `raw/sources/Webinar script.md` — completed the dangling Tools→Memory bridge line, wrote the full **# Memory** section (amnesia demo → notebook tool → close/reopen payoff → system-prompt reveal → user.md beat → tiny-notebook limitation as the Skills hook), and renamed "Markus" → "Marcus" to match the demo app's seeded mail (`Marcus Hale`).
-- Notes: Memory section is grounded in the actual mini-game behavior (Level 04 in `D:\Projects\Names\Webinars\From_chat_to_os`): frozen snapshot per session (hence the mandatory close/reopen), `memory.md` + `user.md` on disk, ~2200-char budget motivating Skills. No wiki pages changed. Next: draft **# Skills** section (open with on-demand loading answering the phone-book limitation).
-
-## 2026-07-14 — maintenance (authorized raw edit, Skills section)
-- Intent: maintenance
-- Input: user request — draft the **# Skills** section in `raw/sources/Webinar script.md` (edit permission granted this session). Corrected flow per user: presenter dictates the operations step by step first, *then* asks AI to package the procedure as a skill.
-- Pages changed: `raw/sources/Webinar script.md` — full Skills section (step-by-step dictation → "I was the recipe" → save-as-skill → SKILL.md on disk → fresh-session one-liner → progressive-disclosure reveal → portability beat → "I'm still the alarm clock" hook into Process).
-- Notes: Grounded in the app's Skills level (Level 05): `save_skill`/`load_skill`, real `skills/<name>/SKILL.md`, live menu (names+descriptions only) rebuilt per window. No wiki pages changed. Next: draft **# Process** (Level 06 — spawn_process scheduler answers the alarm-clock limitation).
-
-## 2026-07-14 — maintenance (authorized raw edit, Process section)
-- Intent: maintenance
-- Input: user request — draft the **# Process** section in `raw/sources/Webinar script.md` (session edit permission).
-- Pages changed: `raw/sources/Webinar script.md` — full Process section (goal not task → spawn_process with pid → hands-off heartbeats → Override slider makes the cube move by itself → ps/kill → reveal: the shell authored the worker's prompt (AI managing AI) → closing arc: same model all six levels, only the harness grew; harness is unique to you; "from a chat box to your own Agentic Operating System"). Also two-branch rule fixes earlier this session (Memory convention + Skills dictation).
-- Notes: Grounded in Level 06 design (`2026-07-11-process-level-design.md`): single ~5s scheduler, stateless ticks, spawn/list/stop syscalls, live −5…+5 weather offset. The "same model, growing harness" close absorbs the `_ideas:` note at the top of the script. Script now has all sections drafted (Intro → Process).
-
-## 2026-07-14 — maintenance (authorized raw edit, OS section)
-- Intent: maintenance
-- Input: user request — add an **# OS** section to `raw/sources/Webinar script.md` (session edit permission): typing removed, mouse-only interaction with the agent.
-- Pages changed: `raw/sources/Webinar script.md` — Process now ends with a "still a terminal / most people never will type" hook; new OS section (one-button app → rule baked in → 5-second checkbox → appliance → "the agent became a program / the chat box dissolved into the OS"); the "same model, growing harness" closing arc moved from Process into OS, with the new final callback "a button that already knows what the email said."
-- Notes: Grounded in the OS app design (`2026-07-12-os-app-design.md`): tile 07, status bar mirrors the agent's closing sentence, shared scheduler slot, Override interplay, rule baked into the prompt. Script ladder now: Chat box → ReAct → Tools → Memory → Skills → Process → OS.
-
-## 2026-07-14 — maintenance (authorized raw edit, OS reframing)
-- Intent: maintenance
-- Input: user correction — the webinar's core idea is "make tools for yourself," not "become a product."
-- Pages changed: `raw/sources/Webinar script.md` — Process→OS hook now ends "…and become a tool? A small one. Made for exactly one person."; OS landing gained the intro callback (weekends burned on internal tools → this one took an evening, asked-for not written) and "It dissolved — into the operating system. Into little tools you make for yourself."; closing arc adds "You don't buy it. You build it — one small tool at a time."
-- Notes: The OS section now closes the loop with the Intro's internal-tools passion. No wiki pages changed.
-
-## 2026-07-14 — ingest (batch, 2 sources: Nina + Yulia interviews)
-- Intent: ingest
-- Input: `raw/sources/Nina interview.md`, `raw/sources/Yulia interview.md` (both are conclusions docs auto-generated by Eugene's record→diarize→transcribe→summarize tool).
-- Pages created — sources (2): [[2026-07-14-nina-interview]], [[2026-07-14-yulia-interview]].
-- Pages created — entities (3): [[nina]] (HR recruiter, Virtido), [[yulia]] (HR lead; name/affiliation tentative, inferred from filename + Nina reference), [[inspectron]] (Eugene's employer).
-- Pages created — concepts (2): [[levels-of-ai-usage]] (chatbot → memory → Claude Code → CLAUDE.md → skills → KB → RAG; non-programmer ceiling = CLAUDE.md + skills), [[solve-first-then-skillify]] (solve once, then freeze into a skill; ~3-message and >5-tool-call heuristics).
-- Pages updated: [[eugene]] (identity evidence strengthened — webinar author/presenter, Inspectron, $200 max plan, tool builder; still tentative), [[virtido]] (HR team = webinar audience, pain points, tooling), [[skills-as-memory]] (handoff/de-risking payoff, ~3-message heuristic, method link), [[connections-as-moat]] (Eugene independently converges: "the human's role is just to be human"), [[overview]] (6 sources, demand-side bullet, agree/diverge, navigation, open questions), `index.md`.
-- Notes: Strong convergence — Eugene ↔ Sebastian/Allie on human-connection residual despite their networking-tactics disagreement; Eugene's ~3-message heuristic ↔ Konstantin's >5-tool-call heuristic. New cross-interview finding: adoption blocked by friction, not resistance. Tentative flags: Yulia's name/affiliation; Eugene's price-rise prediction; Inspectron employment-vs-contracting. Skills-IP-vs-work-product dispute (Eugene vs Sebastian) recorded, unresolved.
-- Next: ingest "Ideas for webinar" / "Webinar Plan" / "Webinar script" to connect corpus → deliverable; interview Larysa/Larisa (PM use cases, twice flagged); consider a comparison page "personal vs employer ownership of skills".
-
-## 2026-07-14 — query (network from a standing start)
-- Intent: query
-- Input: "how does an individual actually build a network from a standing start?" + "go" on the recommended plan (tentative protocol + Sebastian follow-up instrument).
-- Pages created: [[network-from-a-standing-start]] (concept, Status: tentative — 5-step protocol: recurring venues → cadence → lead with humanness → engineer second meetings → track second meetings/let referrals work; contested anti-tactics list; webinar flagged as live case study), [[2026-07-14-network-from-standing-start]] (query — 10-question Sebastian round-2 interview instrument in 3 blocks: bootstrap biography / mechanics / falsification of "Big zero").
-- Pages updated: [[connections-as-moat]] (open question → protocol + validation plan), [[overview]] (vault-level open question annotated), `index.md` (Concepts + Queries sections).
-- Notes: Corpus cannot answer the question directly — protocol is inferred from Sebastian's principles, Allie's prediction, and Eugene's convergent thesis; every claim marked tentative/contested. Key falsification target: whether Sebastian ever had a true standing start (prior-job network as seed). Deep-research complement (weak ties, mere exposure, from-zero playbooks, scoped to test "Big zero") described but not run.
-- Next: run the round-2 Sebastian interview and ingest it; decide on the deep-research run; track webinar → paid-HR-build chain as case-study evidence.
-
-## 2026-07-16 — query → comparison
-- Intent: query
-- Input: "Compare Theo's, Konstantin's ('Sonstantine') and Allie's approaches — what they have in common, what they have different?"
-- Pages created: [[theo-konstantin-allie]] (first `wiki/comparisons/` page — three-lens side-by-side: at-a-glance table, 6 shared points, differences by altitude/framing/code/learning/register, complementary tensions).
-- Pages updated: [[theo-browne]], [[konstantin]], [[allie-miller]] (added inbound "Comparison:" links to avoid orphan); `index.md` (Comparisons section — first entry).
-- Notes: Synthesized from the 3 source summaries + entity pages + [[skills-as-memory]]. Common ground: bottleneck moved model→human system; markdown/skills as the atomic unit; infrastructure compounds; system-ready to absorb new models; context as scarce resource; shared [[claude-code]]. Differences framed as *altitude* (Theo strategy/psychology · Konstantin engineering · Allie individual-productivity), not direction — corpus disagreements lie elsewhere. Recorded the throwaway-code (Theo) vs persist-in-git (Konstantin) tension as complementary (artifact vs capability). No new contradictions. Saved as a comparison rather than a query page since the request was explicitly a side-by-side.
-- Next: reusable starter template (foundation docs / good SKILL.md) still unresolved across all three — webinar-relevant; consider whether Konstantin's meta-loop (wipe-and-restart) tensions with "persist everything".
-
-## 2026-07-16 — maintenance (table fix)
-- Intent: maintenance
-- Input: user report — the table in [[ai-agent-evolution]] "looks broken".
-- Pages changed: [[ai-agent-evolution]] — repaired the "Agent architecture & tooling (Konstantin)" table; escaped the `|` inside aliased wiki links (`[[harness\|harnesses]]`, `[[skills-as-memory\|skills]]`, `[[agentic-loops\|agent loops]]`) so Obsidian no longer reads them as column separators, and removed the two phantom trailing columns the unescaped pipes had forced into the header/separator. Now a clean 3-column table (Period / Milestone / Significance).
-- Notes: Content unchanged — formatting-only fix. The Theo Browne table was already valid. No `index.md` change (no catalog/structural change). Root cause: piped `[[page|alias]]` links inside GFM/Obsidian table cells require the pipe escaped as `\|`.
-
-## 2026-07-21 — ingest (Larysa interview)
-- Intent: ingest
-- Input: "ingest Larysa interview.md" → `raw/sources/Larysa interview.md` (auto-generated interview conclusions doc; Eugene × Larysa, technical BA/PM and ex-mobile dev).
-- Pages created: [[2026-07-21-larysa-interview]] (source), [[larysa]] (entity), [[integration-dead-ends]] (concept — agent begins work against connectors gated by account tier / paid seat / missing API; corpus offers no fix, only up-front verification), [[leave-less-room-for-imagination]] (concept — under-specified prompts drift and the collateral damage is *invisible*; skills as the constraint; includes the 4.7-over-4.8 model-choice corollary).
-- Pages updated: [[skills-as-memory]] (new "negative case" section — built-in memory as anti-feature, the demand-side reason skills exist; contradiction logged vs Allie), [[harness]] (consolidation-over-tool-hopping section + "before and after" claim marked self-reported), [[code-as-throwaway]] (auth/payments trust carve-out; "safe from 4.6 on"), [[levels-of-ai-usage]] (Larysa as proof the rungs are skippable — technically advanced, architecturally stuck), [[claude-code]] (practitioner-limits section: memory, connectors, 4.6/4.7/4.8, emulator), [[solve-first-then-skillify]], [[personal-ai-operating-system]] (cross-ref to Eugene's convergent OS framing), [[eugene]], [[yulia]] (Larisa open question partially resolved), [[virtido]] (audience extends beyond HR; ClickUp/Figma/Teams/Slack), [[overview]] (6→7 sources, demand-side paragraph, two new divergences, two new vault-level open questions), `index.md`.
-- Notes: Closes the "Larysa/Larisa not yet interviewed — PM use cases missing" gap flagged in both HR interviews. Key reframe for the webinar: this source moves the diagnosis from *friction* (Nina/Yulia: people would adopt if it were simple) to *structural walls* for users already past friction — memory, entitlements, drift. Two genuine corpus contradictions recorded rather than smoothed: memory-as-anti-feature (Eugene) vs Allie's untroubled use of persistent context docs, and tight-spec vs [[think-wider-not-bigger]]. Also corrected `index.md`: `HR Contacts.md` was listed as not-yet-ingested but does not exist in `raw/sources/` — removed from the list.
-- Next: remaining un-ingested raw are the three webinar-deliverable docs (`Ideas for webinar.md`, `Webinar Plan - From Chat Box to Your Own OS.md`, `Webinar script.md`) — ingesting them would connect the corpus to the actual deliverable and settle whether the script keeps the [[levels-of-ai-usage]] rung order. Candidate query: does the skills rung actually answer Larysa's memory complaint, or is hers a cross-project problem skills don't solve? A first lint pass is also overdue (no lint reports exist; 7 sources, 17 concepts).
-
-## 2026-07-22 — query (webinar theses)
-- Intent: query
-- Input: "What theses can I suggest for the webinar ('from chatbox to your own agentic operating system') based on what you already have?"
-- Pages changed: created [[2026-07-22-webinar-theses]]; updated `index.md` (Queries section).
-- Notes: 14 candidate theses synthesized from [[overview]], the 7 source summaries, and direct reads of the three raw webinar-deliverable docs (Plan, script, Ideas — still raw-only, cited as raw per policy). Grouped: spine (harness-not-model, app-you-open→OS, skills-as-memory, context scarcity, build-not-buy) / stakes (cost→zero, irreversible gap, connections rise) / obstacles (friction-not-resistance, structural walls, less-room-for-imagination) / method (solve-first-then-skillify, research-vs-judgment, walk-in-a-week) / honest tensions (3 recorded contradictions offered as Q&A ammo). Recommended 7-thesis subset for the 30-min format noted in the query page. No concept/entity pages changed.
-- Next: ingest the three deliverable docs so theses can cite wiki sources instead of raw; decide whether the "obstacles" theses earn a station on the spine.
-
-## 2026-07-22 — ingest (ИИ глупый! short)
-- Intent: ingest
-- Input: "ingest \"ИИ глупый!\"" → `raw/sources/ИИ глупый!.md` (conclusions doc for a 1:28 Russian YouTube Short, author unknown; file was added mid-session after the first lookup found nothing).
-- Pages created: [[2026-07-22-ai-is-stupid]] (source).
-- Pages updated: [[harness]] (business-facing "engineering wrapper" definition + model+context+harness formula, flagged as webinar language), [[context-as-scarce-resource]] (supply-side facet: "intelligence without context loses to context without intelligence", Nobel-vs-employee analogy; RAG-vs-skills tension logged under Contradictions), [[overview]] (7→8 sources, convergence one-liner, navigation), `index.md` (Sources entry; `my theses.md` added to the not-yet-ingested list).
-- Notes: No new concepts — the short *restates* the vault's machine-side spine for a business audience, which makes it webinar-quotable rather than novel. One tension recorded, not smoothed: it recommends RAG/long-term memory as context mechanisms where Konstantin argues skills beat RAG (possibly audience-driven — business data vs procedures). Authorship unknown and marked tentative; its framing matches the corpus closely enough that a connection to someone already in the corpus is worth checking. All ×10…×1000 gain claims are rhetorical, not measured.
-- Next: `my theses.md` is new in raw and un-ingested (six webinar thesis one-liners) — small, high-relevance ingest candidate; the three webinar-deliverable docs remain pending; first lint pass still overdue (8 sources, 17 concepts, 0 lint reports).
-
-## 2026-07-24 — ingest (You're reading way too much code)
-- Intent: ingest
-- Input: "ingest \"You're reading way too much code\"" → `raw/sources/You're reading way too much code.md` (conclusions doc for Theo Browne's 24:11 video; his second source in the vault).
-- Pages created: [[2026-07-24-youre-reading-way-too-much-code]] (source), [[make-more-cheap-code]] (concept — ship/no-ship line, four tiers of code A–D, 100–10,000:1 slop-to-ship verification ratio, always-another-layer, dumb-model agents as API usability testers, reading-costs-attention economics).
-- Pages updated: [[code-as-throwaway]] (discipline section; evidence; its "what is the durable artifact" open question partially answered — the verification harness is now a named first-class output), [[theo-browne]] (second talk, recurring author-move noted, self-reported ratios flagged tentative), [[leave-less-room-for-imagination]] (new logged tension: Theo/Dax agent-diff-summaries vs Eugene's invisible-drift claim), [[2026-07-14-everything-we-knew-about-software-has-changed]] (reciprocal related-source link), [[overview]] (8→9 sources; verifying-half of spine given its method; new divergence line; 17→18 concepts), `index.md` (Sources, People, Concepts entries).
-- Notes: The source is a *defense with discipline*, not a reversal — Theo explicitly rejects shipping unreviewed slop, which pre-empts the obvious objection to [[code-as-throwaway]]. Strong webinar relevance: "attention, not generation, is the bottleneck" is the engineer-side twin of [[context-as-scarce-resource]], and the tier framework gives audiences a non-binary answer to "can I trust AI code?". One genuine tension recorded rather than smoothed (summaries-hide-drift). Shao's 80%-into-harnesses ratio and all of Theo's daily-line counts are self-reported/uncited — marked tentative. Open framework gap flagged: skills/CLAUDE.md files don't fit the A–D spectrum.
-- Next: candidate query — what is the non-engineer's analog of throwaway verification code (HR/BA audience)? `my theses.md` and the three webinar-deliverable docs remain un-ingested; first lint pass still overdue (9 sources, 18 concepts, 0 lint reports).
-
-## 2026-07-24 — query (non-engineer throwaway verification)
-- Intent: query
-- Input: "query: what is the non-engineer's analog of throwaway verification code?" (the candidate query flagged in the previous ingest).
-- Pages created: [[2026-07-24-non-engineer-throwaway-verification]] (query — answer: disposable AI work that attacks/misreads/simulates the deliverable before a human sees it; one-to-one mapping table from Theo's engineer patterns to HR/BA use cases; "checker skills" as the second species of skill; tier-D-stays-human caveat since non-engineer verification bottoms out in judgment, not tests).
-- Pages updated: [[make-more-cheap-code]] (next-question marked answered, link added), [[leave-less-room-for-imagination]] (new next-question: drift harnessed as a sandboxed ambiguity diagnostic — synthesis, tentative), `index.md` (Queries section).
-- Notes: Pure synthesis, no new source — every mapped pattern is grounded in corpus use cases (Nina/Yulia job descriptions, candidate KB, sourcing; Larysa spec ambiguity; Allie's anti-AI-language checker-skill precedent). Key reframes with durable value: (1) the fresh-agent misread test inverts [[leave-less-room-for-imagination]] — drift becomes a diagnostic when sandboxed; (2) [[solve-first-then-skillify]] populates skills in two species, producers and checkers. Flagged honestly: the whole mapping is argument-by-analogy with no measured claim; Nina's transcript-beats-summary finding is a standing counterweight to summary-based review.
-- Next: decide whether "checker skills" earns a slide on the webinar's skills rung; the webinar-deliverable docs (`Ideas`, `Plan`, `script`, `my theses.md`) remain un-ingested; first lint pass still overdue (9 sources, 18 concepts, 2026-07-24 now has 1 query, 0 lint reports).
-
-## 2026-07-28 — sync (script coverage, initial)
-- Intent: sync script coverage
-- Input: first run of Workflow D, establishing `wiki/script-coverage.md`.
-- Pages created: [[script-coverage]] (coverage) — 19 rows, one per concept page.
-- Pages updated: `CLAUDE.md` (folder convention, `#coverage` tag row, Workflow D, new intent), `index.md` (new Coverage section; corrected the not-yet-ingested list — `Ideas for webinar.md` and `my theses.md` were listed under `raw/sources/` but live in `raw/notes/`, and `Agentic Engineering, explained by a 10x developer.md` was missing entirely).
-- Notes: Initial assessment is 5 covered, 4 partial, 10 absent. Eight of the ten absent are human-side or strategy-side per the grouping in index.md; the other two — [[integration-dead-ends]] and [[leave-less-room-for-imagination]] — are machine-side, absent because the script demos the happy path and never reaches connector gating or spec ambiguity. The `ReAct` station carries no wiki concept. No concept pages were edited.
-- Next: decide whether the human-side cluster earns a station or is a deliberate cut; decide separately whether the two absent machine-side failure modes belong in the demo; pin the rows that are decided so future syncs leave them alone.
+# Operation Log
+
+Append-only chronological record of wiki operations. Newest entries at the bottom.
+
+Entry format:
+
+```
+## YYYY-MM-DD — <operation>
+- Intent: ingest | query | lint | maintenance | sync
+- Input: <source path / question / scope>
+- Pages changed: [[...]], [[...]]
+- Notes: <what changed, uncertainty, next steps>
+```
+
+---
+
+## 2026-07-14 — maintenance
+- Intent: maintenance
+- Input: Initialize vault folder structure.
+- Pages changed: created `index.md`, `log.md`, `wiki/overview.md`; created `raw/{sources,assets}` and `wiki/{sources,entities,concepts,timelines,comparisons,queries,lint-reports}`.
+- Notes: Empty vault scaffolded per the LLM Wiki Schema in `CLAUDE.md`. No sources ingested yet. Ready for first "ingest <source>".
+
+## 2026-07-14 — ingest (batch, 4 sources)
+- Intent: ingest
+- Input: The 4 talk/interview docs (user-selected scope; "Ideas for webinar" + "HR Contacts" left as raw reference).
+  - `raw/sources/Everything we knew about software has changed.md`
+  - `raw/sources/In 1 Year, the Gap Between AI Users and Everyone Else Will Be Irreversible.md`
+  - `raw/sources/sebastian interview - conclusions and insights.md`
+  - `raw/sources/Скиллы на базе git — новая память AI-агентов.md`
+- Pages created — sources (4): [[2026-07-14-everything-we-knew-about-software-has-changed]], [[2026-07-14-gap-between-ai-users-irreversible]], [[2026-07-14-sebastian-eugene-interview]], [[2026-07-14-skills-based-on-git]].
+- Pages created — entities (8): [[theo-browne]], [[allie-miller]], [[sebastian]], [[eugene]], [[konstantin]], [[claude-code]], [[hermes]], [[virtido]].
+- Pages created — concepts (13): [[harness]], [[skills-as-memory]], [[evolution-of-agent-tooling]], [[agentic-loops]], [[context-as-scarce-resource]], [[personal-ai-operating-system]], [[product-ownership]], [[connections-as-moat]], [[seniority-and-the-junior-squeeze]], [[decoupling-identity-from-profession]], [[think-wider-not-bigger]], [[code-as-throwaway]], [[enterprise-ai-reality]].
+- Pages created — timelines (1): [[ai-agent-evolution]].
+- Pages updated: [[overview]] (added the cross-source through-line + agree/diverge), `index.md` (full catalog).
+- Notes: Strong cross-source convergence (skills-as-memory: Konstantin ↔ Allie; connections: Sebastian ↔ Allie; code-as-throwaway: Theo ↔ Sebastian). Recorded contradictions (Eugene vs Sebastian on BYO- vs company-managed harness, online vs in-person networking, OSS motivation). Tentative claims flagged: Theo's model-era names, Eugene = vault owner, Hermes/Ralph heuristics, Allie's self-learning-models prediction.
+- Next: consider ingesting "Ideas for webinar" (connects corpus → webinar plan) and "HR Contacts" (the HR-search demo); consider a `wiki/comparisons/` page for the BYO-vs-managed-harness debate; run "lint wiki" to check link health.
+
+## 2026-07-14 — maintenance (tagging convention)
+- Intent: maintenance
+- Input: user request — add per-type hashtags to the wiki schema and existing pages.
+- Pages changed: `CLAUDE.md` (new "Tagging Rules" section, rule 9, templates + lint checklist updated); all 26 wiki pages received a page-type tag line under their H1 — [[overview]] `#overview`, 4 source pages `#source`, 8 entity pages `#entity`, 13 concept pages `#concept`, [[ai-agent-evolution]] `#timeline`; `index.md` (noted the convention).
+- Notes: Tag mapping is folder-driven (`#source`, `#entity`, `#concept`, `#timeline`, `#comparison`, `#query`, `#lint-report`, `#overview`). Exactly one type tag per page, first on the tag line; topical tags may follow. Lint now checks for missing/mismatched type tags.
+
+## 2026-07-14 -- query
+- Intent: query
+- Input: "what would be the best claude skill to give to a claude beginner user?"
+- Pages changed: created [[2026-07-14-best-first-skill-for-beginner]]; updated [[skills-as-memory]] (Next Questions item answered + linked), `index.md` (Queries section).
+- Notes: Answer synthesized from Allie Miller (skill-creator, tone-of-voice/anti-AI-language starter set, foundation-docs-first) and Konstantin (>5-tool-calls auto-creation heuristic). Follow-up: contents of a good tone-of-voice SKILL.md; foundation-doc templates.
+
+## 2026-07-14 — maintenance (authorized raw edit)
+- Intent: maintenance
+- Input: user explicitly authorized editing `raw/sources/Webinar script.md` this session (exception to rule 1).
+- Pages changed: `raw/sources/Webinar script.md` — completed the dangling Tools→Memory bridge line, wrote the full **# Memory** section (amnesia demo → notebook tool → close/reopen payoff → system-prompt reveal → user.md beat → tiny-notebook limitation as the Skills hook), and renamed "Markus" → "Marcus" to match the demo app's seeded mail (`Marcus Hale`).
+- Notes: Memory section is grounded in the actual mini-game behavior (Level 04 in `D:\Projects\Names\Webinars\From_chat_to_os`): frozen snapshot per session (hence the mandatory close/reopen), `memory.md` + `user.md` on disk, ~2200-char budget motivating Skills. No wiki pages changed. Next: draft **# Skills** section (open with on-demand loading answering the phone-book limitation).
+
+## 2026-07-14 — maintenance (authorized raw edit, Skills section)
+- Intent: maintenance
+- Input: user request — draft the **# Skills** section in `raw/sources/Webinar script.md` (edit permission granted this session). Corrected flow per user: presenter dictates the operations step by step first, *then* asks AI to package the procedure as a skill.
+- Pages changed: `raw/sources/Webinar script.md` — full Skills section (step-by-step dictation → "I was the recipe" → save-as-skill → SKILL.md on disk → fresh-session one-liner → progressive-disclosure reveal → portability beat → "I'm still the alarm clock" hook into Process).
+- Notes: Grounded in the app's Skills level (Level 05): `save_skill`/`load_skill`, real `skills/<name>/SKILL.md`, live menu (names+descriptions only) rebuilt per window. No wiki pages changed. Next: draft **# Process** (Level 06 — spawn_process scheduler answers the alarm-clock limitation).
+
+## 2026-07-14 — maintenance (authorized raw edit, Process section)
+- Intent: maintenance
+- Input: user request — draft the **# Process** section in `raw/sources/Webinar script.md` (session edit permission).
+- Pages changed: `raw/sources/Webinar script.md` — full Process section (goal not task → spawn_process with pid → hands-off heartbeats → Override slider makes the cube move by itself → ps/kill → reveal: the shell authored the worker's prompt (AI managing AI) → closing arc: same model all six levels, only the harness grew; harness is unique to you; "from a chat box to your own Agentic Operating System"). Also two-branch rule fixes earlier this session (Memory convention + Skills dictation).
+- Notes: Grounded in Level 06 design (`2026-07-11-process-level-design.md`): single ~5s scheduler, stateless ticks, spawn/list/stop syscalls, live −5…+5 weather offset. The "same model, growing harness" close absorbs the `_ideas:` note at the top of the script. Script now has all sections drafted (Intro → Process).
+
+## 2026-07-14 — maintenance (authorized raw edit, OS section)
+- Intent: maintenance
+- Input: user request — add an **# OS** section to `raw/sources/Webinar script.md` (session edit permission): typing removed, mouse-only interaction with the agent.
+- Pages changed: `raw/sources/Webinar script.md` — Process now ends with a "still a terminal / most people never will type" hook; new OS section (one-button app → rule baked in → 5-second checkbox → appliance → "the agent became a program / the chat box dissolved into the OS"); the "same model, growing harness" closing arc moved from Process into OS, with the new final callback "a button that already knows what the email said."
+- Notes: Grounded in the OS app design (`2026-07-12-os-app-design.md`): tile 07, status bar mirrors the agent's closing sentence, shared scheduler slot, Override interplay, rule baked into the prompt. Script ladder now: Chat box → ReAct → Tools → Memory → Skills → Process → OS.
+
+## 2026-07-14 — maintenance (authorized raw edit, OS reframing)
+- Intent: maintenance
+- Input: user correction — the webinar's core idea is "make tools for yourself," not "become a product."
+- Pages changed: `raw/sources/Webinar script.md` — Process→OS hook now ends "…and become a tool? A small one. Made for exactly one person."; OS landing gained the intro callback (weekends burned on internal tools → this one took an evening, asked-for not written) and "It dissolved — into the operating system. Into little tools you make for yourself."; closing arc adds "You don't buy it. You build it — one small tool at a time."
+- Notes: The OS section now closes the loop with the Intro's internal-tools passion. No wiki pages changed.
+
+## 2026-07-14 — ingest (batch, 2 sources: Nina + Yulia interviews)
+- Intent: ingest
+- Input: `raw/sources/Nina interview.md`, `raw/sources/Yulia interview.md` (both are conclusions docs auto-generated by Eugene's record→diarize→transcribe→summarize tool).
+- Pages created — sources (2): [[2026-07-14-nina-interview]], [[2026-07-14-yulia-interview]].
+- Pages created — entities (3): [[nina]] (HR recruiter, Virtido), [[yulia]] (HR lead; name/affiliation tentative, inferred from filename + Nina reference), [[inspectron]] (Eugene's employer).
+- Pages created — concepts (2): [[levels-of-ai-usage]] (chatbot → memory → Claude Code → CLAUDE.md → skills → KB → RAG; non-programmer ceiling = CLAUDE.md + skills), [[solve-first-then-skillify]] (solve once, then freeze into a skill; ~3-message and >5-tool-call heuristics).
+- Pages updated: [[eugene]] (identity evidence strengthened — webinar author/presenter, Inspectron, $200 max plan, tool builder; still tentative), [[virtido]] (HR team = webinar audience, pain points, tooling), [[skills-as-memory]] (handoff/de-risking payoff, ~3-message heuristic, method link), [[connections-as-moat]] (Eugene independently converges: "the human's role is just to be human"), [[overview]] (6 sources, demand-side bullet, agree/diverge, navigation, open questions), `index.md`.
+- Notes: Strong convergence — Eugene ↔ Sebastian/Allie on human-connection residual despite their networking-tactics disagreement; Eugene's ~3-message heuristic ↔ Konstantin's >5-tool-call heuristic. New cross-interview finding: adoption blocked by friction, not resistance. Tentative flags: Yulia's name/affiliation; Eugene's price-rise prediction; Inspectron employment-vs-contracting. Skills-IP-vs-work-product dispute (Eugene vs Sebastian) recorded, unresolved.
+- Next: ingest "Ideas for webinar" / "Webinar Plan" / "Webinar script" to connect corpus → deliverable; interview Larysa/Larisa (PM use cases, twice flagged); consider a comparison page "personal vs employer ownership of skills".
+
+## 2026-07-14 — query (network from a standing start)
+- Intent: query
+- Input: "how does an individual actually build a network from a standing start?" + "go" on the recommended plan (tentative protocol + Sebastian follow-up instrument).
+- Pages created: [[network-from-a-standing-start]] (concept, Status: tentative — 5-step protocol: recurring venues → cadence → lead with humanness → engineer second meetings → track second meetings/let referrals work; contested anti-tactics list; webinar flagged as live case study), [[2026-07-14-network-from-standing-start]] (query — 10-question Sebastian round-2 interview instrument in 3 blocks: bootstrap biography / mechanics / falsification of "Big zero").
+- Pages updated: [[connections-as-moat]] (open question → protocol + validation plan), [[overview]] (vault-level open question annotated), `index.md` (Concepts + Queries sections).
+- Notes: Corpus cannot answer the question directly — protocol is inferred from Sebastian's principles, Allie's prediction, and Eugene's convergent thesis; every claim marked tentative/contested. Key falsification target: whether Sebastian ever had a true standing start (prior-job network as seed). Deep-research complement (weak ties, mere exposure, from-zero playbooks, scoped to test "Big zero") described but not run.
+- Next: run the round-2 Sebastian interview and ingest it; decide on the deep-research run; track webinar → paid-HR-build chain as case-study evidence.
+
+## 2026-07-16 — query → comparison
+- Intent: query
+- Input: "Compare Theo's, Konstantin's ('Sonstantine') and Allie's approaches — what they have in common, what they have different?"
+- Pages created: [[theo-konstantin-allie]] (first `wiki/comparisons/` page — three-lens side-by-side: at-a-glance table, 6 shared points, differences by altitude/framing/code/learning/register, complementary tensions).
+- Pages updated: [[theo-browne]], [[konstantin]], [[allie-miller]] (added inbound "Comparison:" links to avoid orphan); `index.md` (Comparisons section — first entry).
+- Notes: Synthesized from the 3 source summaries + entity pages + [[skills-as-memory]]. Common ground: bottleneck moved model→human system; markdown/skills as the atomic unit; infrastructure compounds; system-ready to absorb new models; context as scarce resource; shared [[claude-code]]. Differences framed as *altitude* (Theo strategy/psychology · Konstantin engineering · Allie individual-productivity), not direction — corpus disagreements lie elsewhere. Recorded the throwaway-code (Theo) vs persist-in-git (Konstantin) tension as complementary (artifact vs capability). No new contradictions. Saved as a comparison rather than a query page since the request was explicitly a side-by-side.
+- Next: reusable starter template (foundation docs / good SKILL.md) still unresolved across all three — webinar-relevant; consider whether Konstantin's meta-loop (wipe-and-restart) tensions with "persist everything".
+
+## 2026-07-16 — maintenance (table fix)
+- Intent: maintenance
+- Input: user report — the table in [[ai-agent-evolution]] "looks broken".
+- Pages changed: [[ai-agent-evolution]] — repaired the "Agent architecture & tooling (Konstantin)" table; escaped the `|` inside aliased wiki links (`[[harness\|harnesses]]`, `[[skills-as-memory\|skills]]`, `[[agentic-loops\|agent loops]]`) so Obsidian no longer reads them as column separators, and removed the two phantom trailing columns the unescaped pipes had forced into the header/separator. Now a clean 3-column table (Period / Milestone / Significance).
+- Notes: Content unchanged — formatting-only fix. The Theo Browne table was already valid. No `index.md` change (no catalog/structural change). Root cause: piped `[[page|alias]]` links inside GFM/Obsidian table cells require the pipe escaped as `\|`.
+
+## 2026-07-21 — ingest (Larysa interview)
+- Intent: ingest
+- Input: "ingest Larysa interview.md" → `raw/sources/Larysa interview.md` (auto-generated interview conclusions doc; Eugene × Larysa, technical BA/PM and ex-mobile dev).
+- Pages created: [[2026-07-21-larysa-interview]] (source), [[larysa]] (entity), [[integration-dead-ends]] (concept — agent begins work against connectors gated by account tier / paid seat / missing API; corpus offers no fix, only up-front verification), [[leave-less-room-for-imagination]] (concept — under-specified prompts drift and the collateral damage is *invisible*; skills as the constraint; includes the 4.7-over-4.8 model-choice corollary).
+- Pages updated: [[skills-as-memory]] (new "negative case" section — built-in memory as anti-feature, the demand-side reason skills exist; contradiction logged vs Allie), [[harness]] (consolidation-over-tool-hopping section + "before and after" claim marked self-reported), [[code-as-throwaway]] (auth/payments trust carve-out; "safe from 4.6 on"), [[levels-of-ai-usage]] (Larysa as proof the rungs are skippable — technically advanced, architecturally stuck), [[claude-code]] (practitioner-limits section: memory, connectors, 4.6/4.7/4.8, emulator), [[solve-first-then-skillify]], [[personal-ai-operating-system]] (cross-ref to Eugene's convergent OS framing), [[eugene]], [[yulia]] (Larisa open question partially resolved), [[virtido]] (audience extends beyond HR; ClickUp/Figma/Teams/Slack), [[overview]] (6→7 sources, demand-side paragraph, two new divergences, two new vault-level open questions), `index.md`.
+- Notes: Closes the "Larysa/Larisa not yet interviewed — PM use cases missing" gap flagged in both HR interviews. Key reframe for the webinar: this source moves the diagnosis from *friction* (Nina/Yulia: people would adopt if it were simple) to *structural walls* for users already past friction — memory, entitlements, drift. Two genuine corpus contradictions recorded rather than smoothed: memory-as-anti-feature (Eugene) vs Allie's untroubled use of persistent context docs, and tight-spec vs [[think-wider-not-bigger]]. Also corrected `index.md`: `HR Contacts.md` was listed as not-yet-ingested but does not exist in `raw/sources/` — removed from the list.
+- Next: remaining un-ingested raw are the three webinar-deliverable docs (`Ideas for webinar.md`, `Webinar Plan - From Chat Box to Your Own OS.md`, `Webinar script.md`) — ingesting them would connect the corpus to the actual deliverable and settle whether the script keeps the [[levels-of-ai-usage]] rung order. Candidate query: does the skills rung actually answer Larysa's memory complaint, or is hers a cross-project problem skills don't solve? A first lint pass is also overdue (no lint reports exist; 7 sources, 17 concepts).
+
+## 2026-07-22 — query (webinar theses)
+- Intent: query
+- Input: "What theses can I suggest for the webinar ('from chatbox to your own agentic operating system') based on what you already have?"
+- Pages changed: created [[2026-07-22-webinar-theses]]; updated `index.md` (Queries section).
+- Notes: 14 candidate theses synthesized from [[overview]], the 7 source summaries, and direct reads of the three raw webinar-deliverable docs (Plan, script, Ideas — still raw-only, cited as raw per policy). Grouped: spine (harness-not-model, app-you-open→OS, skills-as-memory, context scarcity, build-not-buy) / stakes (cost→zero, irreversible gap, connections rise) / obstacles (friction-not-resistance, structural walls, less-room-for-imagination) / method (solve-first-then-skillify, research-vs-judgment, walk-in-a-week) / honest tensions (3 recorded contradictions offered as Q&A ammo). Recommended 7-thesis subset for the 30-min format noted in the query page. No concept/entity pages changed.
+- Next: ingest the three deliverable docs so theses can cite wiki sources instead of raw; decide whether the "obstacles" theses earn a station on the spine.
+
+## 2026-07-22 — ingest (ИИ глупый! short)
+- Intent: ingest
+- Input: "ingest \"ИИ глупый!\"" → `raw/sources/ИИ глупый!.md` (conclusions doc for a 1:28 Russian YouTube Short, author unknown; file was added mid-session after the first lookup found nothing).
+- Pages created: [[2026-07-22-ai-is-stupid]] (source).
+- Pages updated: [[harness]] (business-facing "engineering wrapper" definition + model+context+harness formula, flagged as webinar language), [[context-as-scarce-resource]] (supply-side facet: "intelligence without context loses to context without intelligence", Nobel-vs-employee analogy; RAG-vs-skills tension logged under Contradictions), [[overview]] (7→8 sources, convergence one-liner, navigation), `index.md` (Sources entry; `my theses.md` added to the not-yet-ingested list).
+- Notes: No new concepts — the short *restates* the vault's machine-side spine for a business audience, which makes it webinar-quotable rather than novel. One tension recorded, not smoothed: it recommends RAG/long-term memory as context mechanisms where Konstantin argues skills beat RAG (possibly audience-driven — business data vs procedures). Authorship unknown and marked tentative; its framing matches the corpus closely enough that a connection to someone already in the corpus is worth checking. All ×10…×1000 gain claims are rhetorical, not measured.
+- Next: `my theses.md` is new in raw and un-ingested (six webinar thesis one-liners) — small, high-relevance ingest candidate; the three webinar-deliverable docs remain pending; first lint pass still overdue (8 sources, 17 concepts, 0 lint reports).
+
+## 2026-07-24 — ingest (You're reading way too much code)
+- Intent: ingest
+- Input: "ingest \"You're reading way too much code\"" → `raw/sources/You're reading way too much code.md` (conclusions doc for Theo Browne's 24:11 video; his second source in the vault).
+- Pages created: [[2026-07-24-youre-reading-way-too-much-code]] (source), [[make-more-cheap-code]] (concept — ship/no-ship line, four tiers of code A–D, 100–10,000:1 slop-to-ship verification ratio, always-another-layer, dumb-model agents as API usability testers, reading-costs-attention economics).
+- Pages updated: [[code-as-throwaway]] (discipline section; evidence; its "what is the durable artifact" open question partially answered — the verification harness is now a named first-class output), [[theo-browne]] (second talk, recurring author-move noted, self-reported ratios flagged tentative), [[leave-less-room-for-imagination]] (new logged tension: Theo/Dax agent-diff-summaries vs Eugene's invisible-drift claim), [[2026-07-14-everything-we-knew-about-software-has-changed]] (reciprocal related-source link), [[overview]] (8→9 sources; verifying-half of spine given its method; new divergence line; 17→18 concepts), `index.md` (Sources, People, Concepts entries).
+- Notes: The source is a *defense with discipline*, not a reversal — Theo explicitly rejects shipping unreviewed slop, which pre-empts the obvious objection to [[code-as-throwaway]]. Strong webinar relevance: "attention, not generation, is the bottleneck" is the engineer-side twin of [[context-as-scarce-resource]], and the tier framework gives audiences a non-binary answer to "can I trust AI code?". One genuine tension recorded rather than smoothed (summaries-hide-drift). Shao's 80%-into-harnesses ratio and all of Theo's daily-line counts are self-reported/uncited — marked tentative. Open framework gap flagged: skills/CLAUDE.md files don't fit the A–D spectrum.
+- Next: candidate query — what is the non-engineer's analog of throwaway verification code (HR/BA audience)? `my theses.md` and the three webinar-deliverable docs remain un-ingested; first lint pass still overdue (9 sources, 18 concepts, 0 lint reports).
+
+## 2026-07-24 — query (non-engineer throwaway verification)
+- Intent: query
+- Input: "query: what is the non-engineer's analog of throwaway verification code?" (the candidate query flagged in the previous ingest).
+- Pages created: [[2026-07-24-non-engineer-throwaway-verification]] (query — answer: disposable AI work that attacks/misreads/simulates the deliverable before a human sees it; one-to-one mapping table from Theo's engineer patterns to HR/BA use cases; "checker skills" as the second species of skill; tier-D-stays-human caveat since non-engineer verification bottoms out in judgment, not tests).
+- Pages updated: [[make-more-cheap-code]] (next-question marked answered, link added), [[leave-less-room-for-imagination]] (new next-question: drift harnessed as a sandboxed ambiguity diagnostic — synthesis, tentative), `index.md` (Queries section).
+- Notes: Pure synthesis, no new source — every mapped pattern is grounded in corpus use cases (Nina/Yulia job descriptions, candidate KB, sourcing; Larysa spec ambiguity; Allie's anti-AI-language checker-skill precedent). Key reframes with durable value: (1) the fresh-agent misread test inverts [[leave-less-room-for-imagination]] — drift becomes a diagnostic when sandboxed; (2) [[solve-first-then-skillify]] populates skills in two species, producers and checkers. Flagged honestly: the whole mapping is argument-by-analogy with no measured claim; Nina's transcript-beats-summary finding is a standing counterweight to summary-based review.
+- Next: decide whether "checker skills" earns a slide on the webinar's skills rung; the webinar-deliverable docs (`Ideas`, `Plan`, `script`, `my theses.md`) remain un-ingested; first lint pass still overdue (9 sources, 18 concepts, 2026-07-24 now has 1 query, 0 lint reports).
+
+## 2026-07-28 — sync (script coverage, initial)
+- Intent: sync script coverage
+- Input: first run of Workflow D, establishing `wiki/script-coverage.md`.
+- Pages created: [[script-coverage]] (coverage) — 19 rows, one per concept page.
+- Pages updated: `CLAUDE.md` (folder convention, `#coverage` tag row, Workflow D, new intent), `index.md` (new Coverage section; corrected the not-yet-ingested list — `Ideas for webinar.md` and `my theses.md` were listed under `raw/sources/` but live in `raw/notes/`, and `Agentic Engineering, explained by a 10x developer.md` was missing entirely).
+- Notes: Initial assessment is 5 covered, 4 partial, 10 absent. Eight of the ten absent are human-side or strategy-side per the grouping in index.md; the other two — [[integration-dead-ends]] and [[leave-less-room-for-imagination]] — are machine-side, absent because the script demos the happy path and never reaches connector gating or spec ambiguity. The `ReAct` station carries no wiki concept. No concept pages were edited.
+- Next: decide whether the human-side cluster earns a station or is a deliberate cut; decide separately whether the two absent machine-side failure modes belong in the demo; pin the rows that are decided so future syncs leave them alone.
diff --git a/wiki/script-coverage.md b/wiki/script-coverage.md
index 1e58577..c707295 100644
--- a/wiki/script-coverage.md
+++ b/wiki/script-coverage.md
@@ -1,45 +1,45 @@
-# Script coverage
-
-#coverage
-
-## Metadata
-
-- **Script:** `raw/sources/Webinar script.md`
-- **Last synced:** 2026-07-28
-- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
-
-## Coverage
-
-| Concept | Status | Station | Pinned |
-|---|---|---|---|
-| [[harness]] | covered | Tools | |
-| [[evolution-of-agent-tooling]] | covered | Tools | |
-| [[skills-as-memory]] | covered | Skills | |
-| [[solve-first-then-skillify]] | covered | Skills | |
-| [[personal-ai-operating-system]] | covered | OS | |
-| [[context-as-scarce-resource]] | partial | Memory | |
-| [[agentic-loops]] | partial | Process | |
-| [[code-as-throwaway]] | partial | OS | |
-| [[levels-of-ai-usage]] | partial | all | |
-| [[integration-dead-ends]] | absent | — | |
-| [[leave-less-room-for-imagination]] | absent | — | |
-| [[product-ownership]] | absent | — | |
-| [[connections-as-moat]] | absent | — | |
-| [[network-from-a-standing-start]] | absent | — | |
-| [[seniority-and-the-junior-squeeze]] | absent | — | |
-| [[decoupling-identity-from-profession]] | absent | — | |
-| [[think-wider-not-bigger]] | absent | — | |
-| [[make-more-cheap-code]] | absent | — | |
-| [[enterprise-ai-reality]] | absent | — | |
-
-## Notes
-
-- Eight of the ten absent concepts are human-side or strategy-side, per the grouping in `index.md`. The script's spine is machine-side and it never reaches that material.
-- The remaining two absent concepts are machine-side: [[integration-dead-ends]] and [[leave-less-room-for-imagination]]. The script demonstrates the happy path, so it never reaches connector gating or spec ambiguity — the two ways the machine side fails in practice.
-- The `ReAct` station carries no wiki concept at all.
-- Set `Pinned` to `yes` on any row whose status is a deliberate decision. Sync will not touch it.
-
-## Related Pages
-
-- [[overview]]
-- `raw/sources/Webinar script.md`
+# Script coverage
+
+#coverage
+
+## Metadata
+
+- **Script:** `raw/sources/Webinar script.md`
+- **Last synced:** 2026-07-28
+- **Stations:** Chat box · ReAct · Tools · Memory · Skills · Process · OS
+
+## Coverage
+
+| Concept | Status | Station | Pinned |
+|---|---|---|---|
+| [[harness]] | covered | Tools | |
+| [[evolution-of-agent-tooling]] | covered | Tools | |
+| [[skills-as-memory]] | covered | Skills | |
+| [[solve-first-then-skillify]] | covered | Skills | |
+| [[personal-ai-operating-system]] | covered | OS | |
+| [[context-as-scarce-resource]] | partial | Memory | |
+| [[agentic-loops]] | partial | Process | |
+| [[code-as-throwaway]] | partial | OS | |
+| [[levels-of-ai-usage]] | partial | all | |
+| [[integration-dead-ends]] | absent | — | |
+| [[leave-less-room-for-imagination]] | absent | — | |
+| [[product-ownership]] | absent | — | |
+| [[connections-as-moat]] | absent | — | |
+| [[network-from-a-standing-start]] | absent | — | |
+| [[seniority-and-the-junior-squeeze]] | absent | — | |
+| [[decoupling-identity-from-profession]] | absent | — | |
+| [[think-wider-not-bigger]] | absent | — | |
+| [[make-more-cheap-code]] | absent | — | |
+| [[enterprise-ai-reality]] | absent | — | |
+
+## Notes
+
+- Eight of the ten absent concepts are human-side or strategy-side, per the grouping in `index.md`. The script's spine is machine-side and it never reaches that material.
+- The remaining two absent concepts are machine-side: [[integration-dead-ends]] and [[leave-less-room-for-imagination]]. The script demonstrates the happy path, so it never reaches connector gating or spec ambiguity — the two ways the machine side fails in practice.
+- The `ReAct` station carries no wiki concept at all.
+- Set `Pinned` to `yes` on any row whose status is a deliberate decision. Sync will not touch it.
+
+## Related Pages
+
+- [[overview]]
+- `raw/sources/Webinar script.md`