token-goat 2.8.0 → 2.8.1

package/README.md

@@ -305,7 +305,7 @@ Installation adds a short routing guide to the agent's instructions. When the ag
 
 | Task | Agent flow |
 | --- | --- |
-| Review a PDF | `pdf-meta` and `pdf-outline`, then `pdf-extract --pages` |
+| Review a PDF | `pdf-meta` and `pdf-outline`, then `pdf-locate` to find the pages that mention a term and `pdf-extract --pages` only those |
 | Review a Word document | `docx-outline`, then `docx-text` |
 | Review a slide deck | `pptx-outline`, then `pptx-slide` or `pptx-notes` |
 | Review a workbook | `xlsx-sheets`, then `xlsx-head`, `xlsx-range`, or `xlsx-query` |
@@ -515,6 +515,7 @@ token-goat xlsx-sheets link-map.xlsx
 token-goat xlsx-query link-map.xlsx --sheet Links --columns publication,source,target --head 50
 token-goat pdf-meta manual.pdf
 token-goat pdf-outline manual.pdf
+token-goat pdf-locate manual.pdf "torque spec" --ignore-case
 token-goat pdf-extract manual.pdf --pages 12-15 --layout --head 120
 ```
 
@@ -606,6 +607,7 @@ token-goat pdf-extract manual.pdf --pages 12-15 --layout --head 120
 | `token-goat config list / get / set / validate` | Inspect or edit `config.toml` from the CLI. `validate` reports unknown keys with did-you-mean suggestions, plus any project-file or environment value that validation rejected or clamped. A project-root `.token-goat.toml` layers on top of the global config, overriding hint thresholds, indexing settings, etc. for that project only. It may not set the security sections `injection`, `webfetch`, `gdrive`, or `mcp`, nor `indexing.cross_project_symbols`: that file arrives with the repository, so a cloned project could otherwise switch off prompt-injection fencing or empty the fetch allow list for anyone who opened it. Those settings come from the global config or the environment only, and a project file that tries to set one is ignored with a message naming what was dropped. `config get`/`list`/`set` report which layer a value actually resolved from — the project file, an environment variable, or the global config — and where that layer's value was clamped or rejected they say so, naming what was asked for and what is in effect instead. |
 | `token-goat config-get <file> <key>` | Look up one key from a config-shaped file (TOML/INI `key = value`, or YAML) without reading the whole thing. On a `.md` file, a leading `---`-fenced YAML frontmatter block (Jekyll/Hugo/SKILL.md style) is checked first and takes precedence over the TOML/INI fallback; a `.md` file with no frontmatter, or an unclosed fence, falls through to the normal lookup unchanged. |
 | `token-goat pdf-extract <file>` | Extract plain text from a PDF instead of a raw Read. `--pages <spec>` narrows to a page range (e.g. `1-5` or `3`); `--head`/`--tail`/`--grep`/`--max-matches`/`--section` slice the extracted text the same way `bash-output`/`web-output` do. `--layout` heuristically reconstructs column-aware reading order from text-item coordinates instead of raw content-stream order (imperfect on rotated/overlapping text). |
+| `token-goat pdf-locate <file> <pattern>` | Find which pages of a PDF match a regex, with a snippet per match, so you can `pdf-extract --pages` only those pages instead of pulling the whole document. `-i`/`--ignore-case` for case-insensitive matching; `--max-matches <n>` caps how many page matches to collect (default 50); `--context <n>` sets the snippet length around each match (default 80); `--pages <spec>` narrows the scan to a page range; `-j`/`--json` emits `{ file, pattern, matchCount, pages, matches }`. |
 | `token-goat pdf-outline <file>` | List a PDF's bookmark/outline tree with page numbers instead of a raw Read. |
 | `token-goat pdf-meta <file> [--json]` | Page count, title/author, and whether a PDF has an extractable text layer (so you know before extracting whether it's scanned/image-only). `--json` emits `{ pageCount, title, author, hasTextLayer }` — `hasTextLayer` as a real boolean rather than a prose sentence, and an absent title/author as `null` rather than the literal `(none)`. |
 | `token-goat image-meta <file> [--json]` | Dimensions, format, byte size, and what a `shrinkImage` pass would cost — a cheap "should I even look at this" probe that reads `sharp` metadata only and never runs OCR. Requires `sharp`; degrades with a clear message when it's missing. |
@@ -643,7 +645,7 @@ token-goat pdf-extract manual.pdf --pages 12-15 --layout --head 120
 | `token-goat zip-list <archive>` | Entry paths and sizes inside a zip-format archive (`.zip`/`.jar`/`.whl`/`.vsix`/`.nupkg` are all zip containers under the hood) instead of a raw Read or an `unzip -l` shell-out. Reads the central directory only — no member is decompressed just to list it. `--json` for structured output. |
 | `token-goat zip-read <archive> <entry>` | Extract and print exactly one entry's text content from a zip-format archive by its in-archive path, instead of extracting the whole archive to disk. A binary member prints a `[binary content elided by token-goat]` marker instead of raw bytes. |
 | `token-goat pr-slice <pr>` | Surgical GitHub PR reads via `gh` — one file's diff, a single review-comment thread, the description, or CI check statuses, instead of pulling the whole PR payload into context. |
-| `token-goat bridges-status` | Parity matrix of which hooks/commands are wired for each supported harness (Claude Code, Codex, opencode, openclaw, Grok, etc.), side by side. |
+| `token-goat bridges-status` | Parity matrix of which hooks/commands are wired for each supported harness (Claude Code, Codex, opencode, openclaw, Grok, etc.), side by side. A `verified` column says how each row was established — `dogfooded` (driven against the real harness binary), `sourced` (read out of the harness's own source or declarations), or `documented` (from its docs only) — so a claim built from reading alone is never presented as one that was tested. |
 | `token-goat commands` | Machine-readable manifest of every registered command, its description, options, and arguments (including subcommands like `worker start`). `--json` emits it as structured JSON for external tooling (shell completion, doc generators, scripts) instead of the default text listing. `--grep PATTERN` narrows the manifest to commands whose name, description, or aliases match; a parent command that matches keeps all its subcommands, a parent that only has a matching child keeps just that child; no matches prints `no matches` and exits 0. |
 | `token-goat mcp-serve` | Run token-goat as an MCP stdio server exposing all 18 tools: read/symbol/section/outline/skeleton/semantic/index_status/refs/brief/map/changed/grep/imports/exports/compress_text/retrieve_text/handoff_create/handoff_resolve. |
 | `token-goat version` | Print the token-goat version. |
@@ -656,7 +658,7 @@ token-goat pdf-extract manual.pdf --pages 12-15 --layout --head 120
 | `token-goat project exclude <path>` | Add a project root to the blocklist so the worker never indexes it. Writes the resolved absolute path to `[worker] blocked_roots` in `config.toml`; idempotent. It also removes anything already indexed under that path and says how many files went, so excluding a directory means its contents stop being readable through `symbol` rather than merely stopping future indexing. Remove the entry from the config to re-enable indexing, then run `token-goat index` to bring the contents back. |
 | `token-goat project prune [--dry-run]` | Remove blocked/excluded roots that no longer exist on disk. `--dry-run` previews removals without touching the config file. Useful after deleting or moving projects. |
 | `token-goat install` | Wire up hooks (and, with the harness flags below, other AI tool integrations). No `--dry-run` or `--verify` flag — run `token-goat doctor` after install to audit the result. |
-| `token-goat doctor` | Confirm everything is wired correctly. Surfaces install state, cold-import timing, cache hit rates, compaction-budget telemetry, opt-in flag status, and canonical-root sanity. A **Security** section reports the posture in one place: whether offline mode is on, whether injection scanning is on, whether the Google Drive integration is enabled, whether fetching runs against an allow list or a deny list, whether MCP reads are confined to the project root, and whether the data directory is readable by other local users. It only warns when a protection that ships on has been switched off, so a default install stays quiet. It read-only audits `~/.copilot/mcp-config.json` for globally configured Chrome DevTools or Playwright `npx` launchers, recommending project scope or removal when inactive; it never prints server configuration or secrets. On Windows, it also reports duplicate Chrome DevTools/Playwright MCP launchers and orphaned Node processes without terminating anything. Pass `--context` to show the **Context footprint** section: a fill bar with severity (ok / warn / high / URGENT), per-component breakdown (skills catalog, loaded skill bodies, CLAUDE.md+MEMORY.md, conversation estimate), session-to-session growth trend with sessions-to-URGENT projection, and tiered compaction recommendations (Tier 0–4) naming the exact commands to run. Auto-shown when fill > 40 % or any loaded skill > 2 K tokens lacks a compact. `--json` emits the check results (one entry per check, with `ok`/`warn`/`fail` status) as JSON instead of text. |
+| `token-goat doctor` | Confirm everything is wired correctly. Surfaces install state, cold-import timing, cache hit rates, compaction-budget telemetry, opt-in flag status, and canonical-root sanity. A **Tool names** check reports any tool name a harness sent that reached no handler wanting it, and calls out the ones that differ from a handled name only by capitalisation or punctuation — the signature of a bridge that forgot to rename something, which is otherwise invisible. A **Security** section reports the posture in one place: whether offline mode is on, whether injection scanning is on, whether the Google Drive integration is enabled, whether fetching runs against an allow list or a deny list, whether MCP reads are confined to the project root, and whether the data directory is readable by other local users. It only warns when a protection that ships on has been switched off, so a default install stays quiet. It read-only audits `~/.copilot/mcp-config.json` for globally configured Chrome DevTools or Playwright `npx` launchers, recommending project scope or removal when inactive; it never prints server configuration or secrets. On Windows, it also reports duplicate Chrome DevTools/Playwright MCP launchers and orphaned Node processes without terminating anything. Pass `--context` to show the **Context footprint** section: a fill bar with severity (ok / warn / high / URGENT), per-component breakdown (skills catalog, loaded skill bodies, CLAUDE.md+MEMORY.md, conversation estimate), session-to-session growth trend with sessions-to-URGENT projection, and tiered compaction recommendations (Tier 0–4) naming the exact commands to run. Auto-shown when fill > 40 % or any loaded skill > 2 K tokens lacks a compact. `--json` emits the check results (one entry per check, with `ok`/`warn`/`fail` status) as JSON instead of text. |
 | `token-goat baseline` | Emit a project map: file count, per-language file counts, the top indexed symbols (by name/kind/location), and the most recently modified files. `--subagent` emits a terser variant (fewer symbols, fewer recent files) for context handed to a freshly spawned subagent; `--json` for the machine-readable form. |
 | `token-goat compact-doc <path>` | Build an extractive compact sidecar for a large reference doc (`.md`/`.markdown`). The compact is stored in the token-goat data dir as a SHA-keyed sidecar; `pre_read` serves it in place of the full file when it exists and is fresh, saving 80–95% of context tokens. Use `--force` to rebuild, `--sentences N` to control lines per section (default 2), `--show` to print the result. The sidecar is automatically marked stale when you edit the source file. Config: `[hints] stable_doc_compacts = true` (default on). |
 

package/dist/{token-goat-chunk-2F6TFBZE.mjs → token-goat-chunk-2G6RAB4G.mjs}

@@ -59,6 +59,7 @@ import {
   propagateEndLinesToSymbols,
   pushAll,
   recordStat,
+  recordUnmappedTool,
   redactIfDotenv,
   redactSecrets,
   resolveIndexPath,
@@ -90,7 +91,7 @@ import {
   withFileLock,
   writeIfDifferent,
   writeJsonSettings
-} from "./token-goat-chunk-44Y77VHR.mjs";
+} from "./token-goat-chunk-ELDJRLHZ.mjs";
 import {
   registerReset
 } from "./token-goat-chunk-AO2QD2AG.mjs";
@@ -2331,10 +2332,18 @@ var parse2 = function(data, opts = {}) {
 function parseRecords(content, opts) {
   const delimiter = opts.delimiter ?? ",";
   if (opts.noHeader === true) {
-    const rows = parse2(content, { columns: false, skip_empty_lines: true, trim: true, delimiter, bom: true });
+    const rows = parse2(content, { columns: false, skip_empty_lines: true, trim: true, delimiter, bom: true, relax_column_count: true });
     return rows.map((row) => Object.fromEntries(row.map((cell, i) => [`col${i + 1}`, cell])));
   }
-  return parse2(content, { columns: true, skip_empty_lines: true, trim: true, delimiter, bom: true });
+  const header = csvHeader(content, opts);
+  const dupes = header.filter((name, i) => name !== "" && header.indexOf(name) !== i);
+  if (dupes.length > 0) {
+    const unique = [...new Set(dupes)];
+    throw new Error(
+      `duplicate column ${unique.length === 1 ? "name" : "names"} in header: ${unique.join(", ")} \u2014 rename the duplicates or pass --no-header to address columns positionally as col1, col2, \u2026`
+    );
+  }
+  return parse2(content, { columns: true, skip_empty_lines: true, trim: true, delimiter, bom: true, relax_column_count: true });
 }
 function csvHeader(content, opts) {
   if (opts.noHeader === true) return [];
@@ -2828,8 +2837,10 @@ function usedRange(ws) {
   }
   const rows = rowCount;
   const cols = maxCol;
-  const ref2 = rows > 0 && cols > 0 ? `A1:${indexToColLetters(cols)}${rows}` : "A1:A1";
-  return { ref: ref2, rows: Math.max(rows, 1), cols: Math.max(cols, 1) };
+  if (rows === 0 || cols === 0) {
+    return { ref: "(empty)", rows: 0, cols: 0 };
+  }
+  return { ref: `A1:${indexToColLetters(cols)}${rows}`, rows, cols };
 }
 async function listSheets(filePath) {
   const wb = await loadWorkbook(filePath);
@@ -2993,6 +3004,40 @@ async function extractPdfText(data, pagesSpec, layout = false) {
     return { text: pages.join("\n\n"), pageCount: doc.numPages, pagesExtracted: end - start + 1 };
   });
 }
+async function locatePdfPages(data, pattern, opts) {
+  let re;
+  try {
+    re = new RegExp(pattern, opts.ignoreCase === true ? "i" : "");
+  } catch (e) {
+    throw new Error(`invalid regex pattern: ${pattern} (${e instanceof Error ? e.message : String(e)})`, { cause: e });
+  }
+  const pdfjs = await loadPdfjs();
+  if (!pdfjs) throw new Error("pdfjs-dist is not installed; run `npm install pdfjs-dist` to enable pdf-extract");
+  const maxMatches = opts.maxMatches ?? 50;
+  const context = opts.context ?? 80;
+  return withPdfDocument(pdfjs, data, async (doc) => {
+    const range = parsePageRange(opts.pages, doc.numPages);
+    const start = range ? range.start : 1;
+    const end = range ? range.end : doc.numPages;
+    const matches = [];
+    for (let i = start; i <= end && matches.length < maxMatches; i++) {
+      const page = await doc.getPage(i);
+      const content = await page.getTextContent();
+      const textItems = content.items.filter((item) => "str" in item);
+      const pageText = textItems.map((item) => item.str).join(" ");
+      const m = re.exec(pageText);
+      if (m === null) continue;
+      matches.push({ page: i, snippet: locateSnippet(pageText, m.index, m[0].length, context) });
+    }
+    return matches;
+  });
+}
+function locateSnippet(text, index, matchLen, context) {
+  const pad = Math.max(0, context - matchLen);
+  const from = Math.max(0, index - Math.floor(pad / 2));
+  const to = Math.min(text.length, index + matchLen + Math.ceil(pad / 2));
+  return text.slice(from, to).replace(/\s+/g, " ").trim();
+}
 async function resolveDestPage(doc, dest) {
   let explicitDest = dest;
   if (typeof explicitDest === "string") {
@@ -3190,9 +3235,27 @@ function toolMatcherFor(eventName) {
   if (parts.length === 0) return null;
   return parts.join("|");
 }
+function foldToolName(name) {
+  return name.toLowerCase().replace(/[_-]/g, "");
+}
+function noteUnrecognizedTool(event, list) {
+  const toolName = event.toolName;
+  if (typeof toolName !== "string" || toolName === "") return;
+  const named = [];
+  for (const { toolName: want } of list) {
+    if (want === void 0) continue;
+    if (want === toolName) return;
+    named.push(want);
+  }
+  if (named.length === 0) return;
+  const folded = foldToolName(toolName);
+  const nearMiss = named.find((n) => foldToolName(n) === folded) ?? null;
+  recordUnmappedTool(toolName, event.eventName, nearMiss);
+}
 async function runHook(event) {
   const list = _handlers.get(event.eventName);
   if (list === void 0) return { hookType: "pass" };
+  noteUnrecognizedTool(event, list);
   let advisoryResult;
   for (const { handler, toolName, advisory } of list) {
     if (toolName !== void 0 && toolName !== event.toolName) continue;
@@ -4237,9 +4300,9 @@ function buildGuidanceBody(fallbackToolClause, opts = {}) {
     "- pulling one value or subtree out of a JSON/YAML/XML file (manifest, lockfile, spec, config) \u2192 `json-query file 'a.b.c'` / `yaml-query file 'a.b.c'` / `xml-query file 'a.b.c'`",
     "- opening an image to check its dimensions, format, or size \u2192 `image-meta file`",
     "- opening a screenshot, diagram, or scan to read the text in it \u2192 `image-text file`",
-    "- opening a PDF or Office document \u2192 inspect its format first, then read a narrow slice: PDF `pdf-meta`/`pdf-outline` then `pdf-extract`; Word `docx-outline` then `docx-text`; PowerPoint `pptx-outline` then `pptx-slide`/`pptx-notes`; Excel `xlsx-sheets` then `xlsx-head`/`xlsx-range`/`xlsx-query`",
+    "- opening a PDF or Office document \u2192 inspect its format first, then read a narrow slice: PDF `pdf-meta`/`pdf-outline` then `pdf-locate` to find the pages and `pdf-extract` only those; Word `docx-outline` then `docx-text`; PowerPoint `pptx-outline` then `pptx-slide`/`pptx-notes`; Excel `xlsx-sheets` then `xlsx-head`/`xlsx-range`/`xlsx-query`",
     "",
-    'Commands: `symbol NAME`, `read "file::symbol"`, `brief "file::symbol"`, `section "file::Heading"`, `semantic "description"`, `outline file`/`skeleton file`, `map --compact`, `refs file::symbol --callers`, `changed --symbol`, `config-get file KEY`, `json-query file \'a.b.c\'`/`yaml-query`/`xml-query`, `json-outline file`/`yaml-outline`/`xml-outline`, `bash-output`/`web-output`/`mcp-output`, ' + (gdrive ? "`gdrive-sections <file-id>`, " : "") + "`image-meta file`/`image-text file`, `pdf-meta`/`pdf-outline`/`pdf-extract`, `docx-outline`/`docx-text`, `pptx-outline`/`pptx-slide`/`pptx-notes`/`pptx-text`, `xlsx-sheets`/`xlsx-head`/`xlsx-range`/`xlsx-query`.",
+    'Commands: `symbol NAME`, `read "file::symbol"`, `brief "file::symbol"`, `section "file::Heading"`, `semantic "description"`, `outline file`/`skeleton file`, `map --compact`, `refs file::symbol --callers`, `changed --symbol`, `config-get file KEY`, `json-query file \'a.b.c\'`/`yaml-query`/`xml-query`, `json-outline file`/`yaml-outline`/`xml-outline`, `bash-output`/`web-output`/`mcp-output`, ' + (gdrive ? "`gdrive-sections <file-id>`, " : "") + "`image-meta file`/`image-text file`, `pdf-meta`/`pdf-outline`/`pdf-locate`/`pdf-extract`, `docx-outline`/`docx-text`, `pptx-outline`/`pptx-slide`/`pptx-notes`/`pptx-text`, `xlsx-sheets`/`xlsx-head`/`xlsx-range`/`xlsx-query`.",
     "",
     "Sub-agent briefs must carry this gate verbatim: a sub-agent inherits none of this context and its reads spend the same token budget.",
     "",
@@ -5107,7 +5170,14 @@ const POLL_ID_ARG_KEY = {
 // the same session, since process.pid varies per invocation -- breaking token-goat's
 // session-based dedup/state ledger, which never accumulates across calls as a result. Derive a
 // stable id instead from the one thing that's actually constant across calls for the same
-// session: the working directory Copilot reports in \`payload.cwd\`.
+// session: the working directory Copilot reports. That field is \`workingDirectory\`, declared
+// required on BaseHookInput in copilot-sdk/types.d.ts since 1.0.76, so it is present on EVERY
+// hook event. This previously read \`payload.cwd\`, which Copilot has never sent under any name in
+// any version -- the key simply did not exist, so this derived every fallback id from
+// process.cwd() instead and \`canonical.cwd\` below was undefined on every single call. It went
+// unnoticed because process.cwd() happens to be the project directory Copilot spawns the hook in,
+// so the fallback was accidentally right; nothing about that was by design. \`cwd\` is still read
+// as a secondary in case a future version adds it under the shorter name.
 function stableFallbackSessionId(cwd) {
   const key = typeof cwd === 'string' && cwd ? cwd : process.cwd()
   const hash = require('node:crypto').createHash('sha256').update(key).digest('hex').slice(0, 16)
@@ -5177,8 +5247,19 @@ async function main() {
 
   const toolName = payload && payload.toolName
   const canonical = {
-    session_id: (payload && payload.sessionId) || stableFallbackSessionId(payload && payload.cwd),
-    cwd: payload && payload.cwd,
+    session_id:
+      (payload && payload.sessionId) ||
+      stableFallbackSessionId(payload && (payload.workingDirectory || payload.cwd)),
+    cwd: payload && (payload.workingDirectory || payload.cwd),
+  }
+
+  // userPromptSubmitted only: Copilot declares \`prompt\` required on UserPromptSubmittedHookInput.
+  // hooks_session.ts's userPromptSubmitHandler reads it as \`event.raw['prompt']\` and gates every
+  // branch it has on the text, so without this it saw '' on every Copilot prompt and the
+  // embedded-skill dedup hint could never fire. Same shape as the postToolUseFailure \`error\`
+  // drop: a required field the canonical builder simply did not list.
+  if (typeof (payload && payload.prompt) === 'string' && payload.prompt !== '') {
+    canonical.prompt = payload.prompt
   }
   if (toolName) {
     canonical.tool_name = TOOL_TO_TG[toolName] || toolName
@@ -5193,6 +5274,16 @@ async function main() {
   // post-read/post-bash stats stay empty no matter how many tool calls happen. Extract the
   // LLM-facing text directly rather than forwarding the raw object, since textResultForLlm
   // isn't one of those recognized object keys.
+  // postToolUseFailure only: Copilot's PostToolUseFailureHookInput (copilot-sdk/types.d.ts:1042)
+  // carries {toolName, toolArgs, error} and no toolResult at all -- the failure text lives in
+  // \`error\`, a plain string. Without forwarding it, hooks_tool_failure.ts's extractFailureText
+  // finds nothing to key on and the repeat-failure brake returns pass on every single call: wired,
+  // green, and doing nothing. Found by driving the installed shim rather than by a test, because
+  // the handler's own tests hand it a raw payload that already has the field.
+  if (typeof (payload && payload.error) === 'string' && payload.error !== '') {
+    canonical.error = payload.error
+  }
+
   const rawResult = payload && payload.toolResult
   if (rawResult && typeof rawResult === 'object') {
     const tr = rawResult
@@ -7373,6 +7464,9 @@ function resolveTesseractEntry() {
   }
   return _tesseractEntryPath;
 }
+function isOcrEngineAvailable() {
+  return !_ocrUnavailableThisProcess && resolveTesseractEntry() !== null;
+}
 var _ocrUnavailableThisProcess = false;
 function buildChildScript(entryPath, cacheDir) {
   return [
@@ -7487,14 +7581,16 @@ var loadSharp = createLazyModuleLoader(async () => {
 function isImagePath(p) {
   return IMAGE_EXTENSIONS.has(path12.extname(p).toLowerCase());
 }
+var ImageDecodeError = class extends Error {
+};
 async function probeImageMeta(input) {
   const sharp = await loadSharp();
   if (sharp === null) return null;
   try {
     const meta = await sharp(input, { limitInputPixels: false }).metadata();
     return { width: meta.width ?? 0, height: meta.height ?? 0, format: meta.format ?? null, pages: meta.pages ?? 1 };
-  } catch {
-    return null;
+  } catch (e) {
+    throw new ImageDecodeError(e?.message ?? "image could not be decoded");
   }
 }
 async function imageQualifiesForShrink(input) {
@@ -7640,7 +7736,14 @@ async function preReadImageHandler(event) {
     } catch {
       cachedData = null;
     }
-    const meta = cachedData !== null ? await probeImageMeta(cachedData) : null;
+    let meta = null;
+    if (cachedData !== null) {
+      try {
+        meta = await probeImageMeta(cachedData);
+      } catch {
+        meta = null;
+      }
+    }
     if (cachedData !== null && meta !== null) {
       const result2 = {
         data: cachedData,
@@ -18839,6 +18942,7 @@ function urlPolicyDenialReason(url, policy) {
 export {
   createLazyModuleLoader,
   extractPdfText,
+  locatePdfPages,
   extractPdfOutline,
   extractPdfMeta,
   docxOutline,
@@ -19005,10 +19109,12 @@ export {
   getSkillFilePath,
   installedSkillPath,
   pruneSkillOutputs,
+  isOcrEngineAvailable,
   ocrImage,
   isTextHeavy,
   formatShrinkSummary,
   isImagePath,
+  ImageDecodeError,
   probeImageMeta,
   imageQualifiesForShrink,
   shrinkImage,

package/dist/{token-goat-chunk-FBGBTICM.mjs → token-goat-chunk-73MF6YWW.mjs}

@@ -25,7 +25,7 @@ import {
   runSkeleton,
   runSymbol,
   withPinnedReads
-} from "./token-goat-chunk-SRAR6DOK.mjs";
+} from "./token-goat-chunk-RMDQFTQD.mjs";
 import {
   buildProjectMap,
   embeddingsDepsAvailable,
@@ -34,7 +34,7 @@ import {
   getProjectIndexCounts,
   isWorkerRunning,
   mapLookupBytesSaved
-} from "./token-goat-chunk-2F6TFBZE.mjs";
+} from "./token-goat-chunk-2G6RAB4G.mjs";
 import {
   VERSION,
   dataDir,
@@ -45,7 +45,7 @@ import {
   normalizePath,
   recordStat,
   resolveProjectRoot
-} from "./token-goat-chunk-44Y77VHR.mjs";
+} from "./token-goat-chunk-ELDJRLHZ.mjs";
 import "./token-goat-chunk-AO2QD2AG.mjs";
 import "./token-goat-chunk-AEX54RUZ.mjs";
 

package/dist/{token-goat-chunk-TUPJRK7R.mjs → token-goat-chunk-7JDXDERZ.mjs}

@@ -4,7 +4,7 @@ import {
   loadConfig,
   redactSecrets,
   stripAnsiCodes
-} from "./token-goat-chunk-44Y77VHR.mjs";
+} from "./token-goat-chunk-ELDJRLHZ.mjs";
 
 // src/tool_filters/helpers.ts
 import * as fs from "node:fs";

package/dist/{token-goat-chunk-LVCBDJVE.mjs → token-goat-chunk-AOBZUFNJ.mjs}

@@ -23,7 +23,7 @@ import {
   summarizeOutputDelta,
   summarizeResidentContext,
   taskListPruneHint
-} from "./token-goat-chunk-AO6MFFTW.mjs";
+} from "./token-goat-chunk-XEDQH5DA.mjs";
 import {
   BODY_FIRST_TOOL_RESPONSE_KEYS,
   OUTPUT_FIRST_TOOL_RESPONSE_KEYS,
@@ -118,7 +118,7 @@ import {
   wasCliReadThisSession,
   wasFileReadThisSession,
   wasHintShown
-} from "./token-goat-chunk-2F6TFBZE.mjs";
+} from "./token-goat-chunk-2G6RAB4G.mjs";
 import {
   canRunWrappedShell,
   compressOutput,
@@ -129,7 +129,7 @@ import {
   isRewriteWorthwhile,
   resolveMinNetSavingsBytes,
   shlexSplit
-} from "./token-goat-chunk-TUPJRK7R.mjs";
+} from "./token-goat-chunk-7JDXDERZ.mjs";
 import {
   VERSION,
   detectHarness,
@@ -149,7 +149,7 @@ import {
   runGit,
   shortFingerprint,
   toKB
-} from "./token-goat-chunk-44Y77VHR.mjs";
+} from "./token-goat-chunk-ELDJRLHZ.mjs";
 
 // src/hooks_grep.ts
 function grepIntInput(toolInput, key) {
@@ -620,7 +620,7 @@ import crypto from "node:crypto";
 var TRACKED_SKILL = "token-goat";
 var MAX_COMMANDS_SHOWN = 8;
 async function currentCommandNames() {
-  const { buildProgram } = await import("./token-goat-chunk-J35AKWEQ.mjs");
+  const { buildProgram } = await import("./token-goat-chunk-Y2SYNH3S.mjs");
   return flattenCommandNames(buildCommandManifest(buildProgram()));
 }
 async function recordSkillVersionSnapshot(sessionId, skillName) {

package/dist/{token-goat-chunk-44Y77VHR.mjs → token-goat-chunk-ELDJRLHZ.mjs}

@@ -8,7 +8,7 @@ import {
 import { createRequire } from "node:module";
 function resolveVersion() {
   if (true) {
-    return "2.8.0";
+    return "2.8.1";
   }
   const require2 = createRequire(import.meta.url);
   const pkg = require2("../package.json");
@@ -5801,6 +5801,30 @@ function _renderByCommandSection(stats) {
   }
   return lines;
 }
+function _renderByHarnessSection(stats) {
+  if (!stats.by_harness || stats.by_harness.length < 2) {
+    return [];
+  }
+  const lines = [..._sectionHeader("By harness"), _tableHeader("harness")];
+  const { grossBytes, shareBytesDenom, shareTokensDenom } = _computeShareDenominators(stats.by_harness);
+  function share(h) {
+    return _absShare(h.bytes, h.tokens, shareBytesDenom, shareTokensDenom);
+  }
+  for (const h of [...stats.by_harness].sort((a, b) => share(b) - share(a))) {
+    lines.push(
+      _tableRow({
+        name: h.harness,
+        fraction: _barFraction(h.bytes, grossBytes),
+        bytes: h.bytes,
+        tokens: h.tokens,
+        events: h.events,
+        share: share(h),
+        nameColor: C.TEXT_PRIMARY
+      })
+    );
+  }
+  return lines;
+}
 function _renderByDaySection(stats) {
   if (stats.by_day.length === 0) {
     return [];
@@ -5923,6 +5947,7 @@ function renderStats(stats, opts) {
     _renderByKindSection(stats),
     _renderBySourceSection(stats),
     _renderByCommandSection(stats),
+    _renderByHarnessSection(stats),
     _renderByDaySection(stats),
     _renderByProjectSection(stats),
     _renderInsightsSection(stats),
@@ -5932,6 +5957,7 @@ function renderStats(stats, opts) {
 }
 
 // src/stats.ts
+var HARNESS_UNRECORDED = "unrecorded (pre-2.8.1)";
 var SOURCE_IMAGE = "image";
 var SOURCE_HINT = "hint";
 var SOURCE_READ = "read";
@@ -5978,6 +6004,7 @@ var KIND_TO_SOURCE = {
   csv_query: SOURCE_READ,
   csv_profile: SOURCE_READ,
   pdf_extract: SOURCE_READ,
+  pdf_locate: SOURCE_READ,
   pdf_outline: SOURCE_READ,
   pdf_meta: SOURCE_READ,
   xlsx_sheets: SOURCE_READ,
@@ -6067,6 +6094,7 @@ var COMMAND_KINDS = {
   "csv-query": /* @__PURE__ */ new Set(["csv_query"]),
   "csv-profile": /* @__PURE__ */ new Set(["csv_profile"]),
   "pdf-extract": /* @__PURE__ */ new Set(["pdf_extract"]),
+  "pdf-locate": /* @__PURE__ */ new Set(["pdf_locate"]),
   "pdf-outline": /* @__PURE__ */ new Set(["pdf_outline"]),
   "pdf-meta": /* @__PURE__ */ new Set(["pdf_meta"]),
   "xlsx-sheets": /* @__PURE__ */ new Set(["xlsx_sheets"]),
@@ -6161,19 +6189,53 @@ CREATE TABLE IF NOT EXISTS stats (
   kind TEXT NOT NULL,
   tokens_saved INTEGER NOT NULL DEFAULT 0,
   bytes_saved INTEGER NOT NULL DEFAULT 0,
-  detail TEXT
+  detail TEXT,
+  harness TEXT
 );
 CREATE INDEX IF NOT EXISTS idx_stats_ts ON stats(ts);
 CREATE INDEX IF NOT EXISTS idx_stats_kind ON stats(kind);
+CREATE TABLE IF NOT EXISTS unmapped_tools (
+  harness TEXT NOT NULL,
+  tool_name TEXT NOT NULL,
+  event_name TEXT NOT NULL,
+  near_miss TEXT,
+  first_seen INTEGER NOT NULL,
+  last_seen INTEGER NOT NULL,
+  hits INTEGER NOT NULL DEFAULT 0,
+  PRIMARY KEY (harness, tool_name, event_name)
+);
 `;
 var _globalSchemaApplied = /* @__PURE__ */ new Set();
 registerReset(() => _globalSchemaApplied.clear());
+function migrateGlobalSchema(db) {
+  try {
+    db.exec("ALTER TABLE stats ADD COLUMN harness TEXT");
+  } catch (err) {
+    if (!(err instanceof Error) || !/duplicate column/i.test(err.message)) throw err;
+  }
+}
+var _harnessColumnByDb = /* @__PURE__ */ new WeakMap();
+function statsHasHarnessColumn(db) {
+  const cached = _harnessColumnByDb.get(db);
+  if (cached !== void 0) return cached;
+  let present;
+  try {
+    present = db.prepare("PRAGMA table_info(stats)").all().some(
+      (c) => c.name === "harness"
+    );
+  } catch {
+    present = false;
+  }
+  _harnessColumnByDb.set(db, present);
+  return present;
+}
 function getGlobalDb(homeDir) {
   const basePath = homeDir ? dataDirForHome(homeDir) : dataDir();
   const dbPath = path8.join(basePath, "global.db");
   const db = getDb(dbPath);
   if (!_globalSchemaApplied.has(dbPath)) {
     db.exec(GLOBAL_SCHEMA_SQL);
+    migrateGlobalSchema(db);
     _globalSchemaApplied.add(dbPath);
   }
   return db;
@@ -6188,10 +6250,44 @@ function noStatsMessage(windowDays, homeDir) {
 function recordStat(kind, bytesSaved = 0, tokensSaved = 0, _testDb, detail) {
   try {
     const db = _testDb ?? getGlobalDb();
+    const ts = Math.floor(Date.now() / 1e3);
+    if (statsHasHarnessColumn(db)) {
+      db.prepare(
+        "INSERT INTO stats (ts, kind, bytes_saved, tokens_saved, detail, harness) VALUES (?, ?, ?, ?, ?, ?)"
+      ).run(ts, kind, bytesSaved, tokensSaved, detail ?? null, getHarnessName());
+    } else {
+      db.prepare(
+        "INSERT INTO stats (ts, kind, bytes_saved, tokens_saved, detail) VALUES (?, ?, ?, ?, ?)"
+      ).run(ts, kind, bytesSaved, tokensSaved, detail ?? null);
+    }
+  } catch {
+  }
+}
+var MAX_TOOL_NAME_CHARS = 200;
+function recordUnmappedTool(toolName, eventName, nearMiss, _testDb) {
+  try {
+    if (!toolName) return;
+    const db = _testDb ?? getGlobalDb();
+    const now = Math.floor(Date.now() / 1e3);
     db.prepare(
-      "INSERT INTO stats (ts, kind, bytes_saved, tokens_saved, detail) VALUES (?, ?, ?, ?, ?)"
-    ).run(Math.floor(Date.now() / 1e3), kind, bytesSaved, tokensSaved, detail ?? null);
+      `INSERT INTO unmapped_tools (harness, tool_name, event_name, near_miss, first_seen, last_seen, hits)
+       VALUES (?, ?, ?, ?, ?, ?, 1)
+       ON CONFLICT(harness, tool_name, event_name) DO UPDATE SET
+         hits = hits + 1,
+         last_seen = excluded.last_seen,
+         near_miss = excluded.near_miss`
+    ).run(getHarnessName(), toolName.slice(0, MAX_TOOL_NAME_CHARS), eventName, nearMiss, now, now);
+  } catch {
+  }
+}
+function readUnmappedTools(dbPath, homeDir) {
+  try {
+    const db = dbPath ? getDb(dbPath) : getGlobalDb(homeDir);
+    return db.prepare(
+      "SELECT harness, tool_name, event_name, near_miss, hits, last_seen FROM unmapped_tools ORDER BY hits DESC, tool_name ASC"
+    ).all();
   } catch {
+    return [];
   }
 }
 function summarize(windowDays = 30, testDb, homeDir) {
@@ -6199,11 +6295,14 @@ function summarize(windowDays = 30, testDb, homeDir) {
   const sinceTs = windowDays > 0 ? Math.floor((Date.now() - windowDays * 24 * 60 * 60 * 1e3) / 1e3) : null;
   const byKind = {};
   const byDay = {};
+  const byHarness = {};
   let totalEvents = 0;
   let totalBytes = 0;
   let totalTokens = 0;
   const db = testDb ?? getGlobalDb(homeDir);
-  const query = sinceTs !== null ? "SELECT ts, kind, bytes_saved, tokens_saved FROM stats WHERE ts >= ? ORDER BY ts DESC" : "SELECT ts, kind, bytes_saved, tokens_saved FROM stats ORDER BY ts DESC";
+  const hasHarness = statsHasHarnessColumn(db);
+  const cols = hasHarness ? "ts, kind, bytes_saved, tokens_saved, harness" : "ts, kind, bytes_saved, tokens_saved";
+  const query = sinceTs !== null ? `SELECT ${cols} FROM stats WHERE ts >= ? ORDER BY ts DESC` : `SELECT ${cols} FROM stats ORDER BY ts DESC`;
   const stmt = db.prepare(query);
   const rows = sinceTs !== null ? stmt.all(sinceTs) : stmt.all();
   const tsToDateCache = {};
@@ -6227,6 +6326,11 @@ function summarize(windowDays = 30, testDb, homeDir) {
       byDay[dateKey] = zeroBucket();
     }
     incBucket(byDay[dateKey], bytesSaved, tokensSaved);
+    const harness = row.harness || HARNESS_UNRECORDED;
+    if (!byHarness[harness]) {
+      byHarness[harness] = zeroBucket();
+    }
+    incBucket(byHarness[harness], bytesSaved, tokensSaved);
   }
   const bySourceDict = {};
   for (const [kind, bucket] of Object.entries(byKind)) {
@@ -6263,6 +6367,7 @@ function summarize(windowDays = 30, testDb, homeDir) {
     by_day: byDayList,
     by_project: byProjectList,
     by_source: bySourceDict,
+    by_harness: byHarness,
     by_command: Object.entries(byCommandDict).map(([command, bucket]) => ({ ...bucket, command })).filter((r) => r.events > 0),
     window_days: windowDays
   };
@@ -6300,6 +6405,15 @@ function _plainTextStats(summary) {
       );
     }
   }
+  const harnesses = Object.entries(summary.by_harness).filter(([, b]) => b.events > 0).sort((a, b) => b[1].tokens_saved - a[1].tokens_saved);
+  if (harnesses.length > 1) {
+    lines.push("", "## By Harness");
+    for (const [harness, bucket] of harnesses) {
+      lines.push(
+        `  ${harness.padEnd(22)} ${bucket.events.toString().padStart(6)} events  ${fmtBytes(bucket.bytes_saved).padStart(8)}  ${bucket.tokens_saved.toString().padStart(8)} tokens`
+      );
+    }
+  }
   if (summary.by_command.length > 0) {
     lines.push("", "## By Command");
     for (const row of summary.by_command) {
@@ -6373,7 +6487,13 @@ function _buildStatsData(summary, windowDays) {
       bytes: c.bytes_saved,
       tokens: c.tokens_saved,
       events: c.events
-    }))
+    })),
+    by_harness: Object.entries(summary.by_harness).filter(([, b]) => b.events > 0).map(([harness, bucket]) => ({
+      harness,
+      bytes: bucket.bytes_saved,
+      tokens: bucket.tokens_saved,
+      events: bucket.events
+    })).sort((a, b) => b.bytes - a.bytes)
   };
 }
 function renderShortStats(opts) {
@@ -6770,6 +6890,8 @@ export {
   SOURCE_HINT,
   formatLocalTimestamp,
   recordStat,
+  recordUnmappedTool,
+  readUnmappedTools,
   summarize,
   _useRichStats,
   renderShortStats,

package/dist/{token-goat-chunk-4KZILRZN.mjs → token-goat-chunk-I6TOUPLP.mjs}

@@ -10,11 +10,11 @@ import {
   selectFilter,
   shlexSplit,
   wrappedShell
-} from "./token-goat-chunk-TUPJRK7R.mjs";
+} from "./token-goat-chunk-7JDXDERZ.mjs";
 import {
   loadConfig,
   recordStat
-} from "./token-goat-chunk-44Y77VHR.mjs";
+} from "./token-goat-chunk-ELDJRLHZ.mjs";
 import "./token-goat-chunk-AO2QD2AG.mjs";
 import "./token-goat-chunk-AEX54RUZ.mjs";