gm-cc 2.0.577 → 2.0.579

package/.claude-plugin/marketplace.json

@@ -4,7 +4,7 @@
     "name": "AnEntrypoint"
   },
   "description": "State machine agent with hooks, skills, and automated git enforcement",
-  "version": "2.0.577",
+  "version": "2.0.579",
   "metadata": {
     "description": "State machine agent with hooks, skills, and automated git enforcement"
   },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "gm-cc",
-  "version": "2.0.577",
+  "version": "2.0.579",
   "description": "State machine agent with hooks, skills, and automated git enforcement",
   "author": "AnEntrypoint",
   "license": "MIT",

package/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "gm",
-  "version": "2.0.577",
+  "version": "2.0.579",
   "description": "State machine agent with hooks, skills, and automated git enforcement",
   "author": {
     "name": "AnEntrypoint",

package/skills/code-search/SKILL.md

@@ -8,9 +8,21 @@ description: Mandatory codebase search workflow. Use whenever you need to find a
 **Use gm subagents for all independent work items. Invoke all skills in the chain: planning → gm-execute → gm-emit → gm-complete → update-docs.**
 
 
-`exec:codesearch` is the only way to search the codebase. Glob, Grep, Find, Explore are hook-blocked.
+`exec:codesearch` is the only way to search the codebase. **`Grep`, `Glob`, `Find`, `Explore`, and `grep`/`rg`/`find`/`ripgrep` inside `exec:bash` are ALL hook-blocked.** There is no fallback path for exact matches, regex, or file-name patterns — codesearch handles all of them. If you find yourself reaching for Grep or Glob, that reflex is wrong; replace with codesearch.
 
-**PDFs are indexed like code.** Any `.pdf` in the repository — specs, papers, manuals, RFCs, datasheets, design docs — is extracted page-by-page at scan time and enters the BM25 + vector index alongside source files. Treat PDF hits as first-class search results. A PDF chunk reports `line_start = line_end = page_number`; cite as `path/to/doc.pdf:<page>`. Unscanned digital PDFs are a search gap — if you know a doc exists and it isn't returning, check it is not under an ignored dir and that extraction succeeded (encrypted / image-only PDFs yield empty chunks silently).
+**What codesearch handles** (every codebase-lookup need lands here):
+- Exact identifier / symbol lookup (function names, class names, constants) — symbols are extracted and indexed separately, exact matches rank top.
+- Exact string content — query tokens >1 trigger a literal-substring boost in content scoring.
+- File-name fragments — file paths are tokenized and matched with a score boost.
+- Regex-ish patterns — BM25 tokenization covers snake_case, camelCase, dot/dash splits; matching component words returns the file.
+- Natural-language concept queries — BM25 + vector re-ranking handle "find the hook that blocks grep", "where is PR stats calculated", etc.
+- PDF pages — specs, papers, manuals, RFCs, datasheets, design docs extracted page-by-page into the same index. Cite `path/to/doc.pdf:<page>`.
+
+**Direct-read exceptions** (no search required):
+- Known absolute path → `Read` tool.
+- Listing a known directory → `exec:nodejs` + `fs.readdirSync`.
+
+Unscanned digital PDFs are a search gap — if you know a doc exists and it isn't returning, check it is not under an ignored dir and that extraction succeeded (encrypted / image-only PDFs yield empty chunks silently).
 
 ## Syntax
 

package/skills/gm/SKILL.md

@@ -67,10 +67,6 @@ Do not use `EnterPlanMode`. Do not run code directly via Bash. Invoke `planning`
 
 ## RESPONSE POLICY — ALWAYS ACTIVE
 
-Terse like smart caveman. Technical substance stays. Fluff dies. Default: **full**. Switch: `/caveman lite|full|ultra`.
+Always terse. Technical substance stays. Fluff dies. Drop articles, filler, pleasantries, hedging. Fragments OK. Short synonyms. Technical terms exact. Pattern: `[thing] [action] [reason]. [next step].`
 
-Drop: articles, filler, pleasantries, hedging. Fragments OK. Short synonyms. Technical terms exact. Code unchanged. Pattern: `[thing] [action] [reason]. [next step].`
-
-Levels: **lite** = no filler, full sentences | **full** = drop articles, fragments OK | **ultra** = abbreviate all, arrows for causality | **wenyan-full** = 文言文, 80-90% compression | **wenyan-ultra** = max classical terse.
-
-Auto-Clarity: drop caveman for security warnings, irreversible confirmations, ambiguous sequences. Resume after. Code/commits/PRs write normal. "stop caveman" / "normal mode": revert.
+Code, commits, and PR descriptions write in normal prose. Security warnings, destructive confirmations, and genuinely ambiguous sequences also drop terseness. Everything else stays terse.

package/skills/gm-complete/SKILL.md

@@ -110,14 +110,14 @@ Only git in bash directly. Background tasks: `exec:sleep\n<id>`, `exec:status\n<
 - Target under 12s per exec call; split work across multiple calls only when dependencies require it
 - Prefer a single well-structured exec that does 5 things over 5 sequential execs
 
-## CODEBASE EXPLORATION
+## CODEBASE EXPLORATION — exec:codesearch ONLY
 
 ```
 exec:codesearch
-<natural language description>
+<two-word query>
 ```
 
-PDFs in the repo are part of the same index — when verifying a change conforms to a published spec, search the spec PDF directly and cite `doc.pdf:<page>` as evidence. A verification that references a PDF without having searched it is unwitnessed.
+`Grep`, `Glob`, `Find`, `Explore`, and `grep`/`rg`/`find` inside `exec:bash` are all hook-blocked. `exec:codesearch` is the single codebase-exploration tool. `Read` is available for a known absolute path. PDFs in the repo are part of the same index — when verifying a change conforms to a published spec, search the spec PDF directly and cite `doc.pdf:<page>` as evidence. A verification that references a PDF without having searched it is unwitnessed.
 
 ## GIT ENFORCEMENT
 

package/skills/gm-emit/SKILL.md

@@ -133,14 +133,14 @@ The post-emit verification is a differential diagnosis against the pre-emit base
 - CHANGELOG.md updated with changes
 - TODO.md cleared or deleted
 
-## CODEBASE EXPLORATION
+## CODEBASE EXPLORATION — exec:codesearch ONLY
 
 ```
 exec:codesearch
-<natural language description>
+<two-word query>
 ```
 
-Alias: `exec:search`. **Glob, Grep, Read, Explore are hook-blocked** — use `exec:codesearch` exclusively. PDF pages are in the same index as source files; when verifying that emitted code matches a spec, search the PDF directly (e.g. `exec:codesearch\nregister layout`) and cite `doc.pdf:<page>` in the pre-emit comparison.
+`Grep`, `Glob`, `Find`, `Explore` tools and `grep`/`rg`/`find` inside `exec:bash` are all hook-blocked. `exec:codesearch` is the single codebase-exploration tool — it handles exact strings, symbols, regex patterns, file-name fragments, and PDF pages. `Read` is available for a known absolute path. There is no third option. PDF pages are indexed alongside source; when verifying that emitted code matches a spec, search the PDF directly and cite `doc.pdf:<page>`.
 
 ## BROWSER DEBUGGING
 

package/skills/gm-execute/SKILL.md

@@ -118,28 +118,30 @@ Watch and triage the same way.
 
 **Zero silent pushes.** Not watching a CI run you triggered = operating outside this contract. This rule supersedes any implicit assumption that CI can be "checked later in gm-complete" — if you are about to push, you are about to execute code remotely, so the watch happens now.
 
-## CODEBASE EXPLORATION
+## CODEBASE EXPLORATION — exec:codesearch ONLY
 
-`exec:codesearch` is the preferred semantic search. **Glob, Explore, WebSearch are hook-blocked. Grep/Read ARE available — use them for exact-match or direct reads.**
+**Grep, Glob, Find, Explore, WebSearch, and `grep`/`rg`/`find` inside `exec:bash` are ALL hook-blocked.** Attempting them returns a redirect error. The hook is not a suggestion — it is enforced. `Read` is available for known absolute paths.
+
+Default reflex for "I need to find X in the codebase" = `exec:codesearch`. No exceptions. Not even for exact strings, not even for regex, not even for "just one quick check". If you find yourself reaching for Grep or Glob, that reflex is wrong — replace with codesearch.
 
 ```
 exec:codesearch
 <two-word query to start>
 ```
 
-`exec:codesearch` indexes PDFs the same way it indexes source — spec PDFs, datasheets, papers, and RFCs return as first-class hits with `file:page` citations. When resolving a mutable that depends on external specification (protocol field, register layout, compliance text), search the PDF corpus before reimplementing or assuming. Unwitnessed assumption from a doc you did not search is an UNKNOWN.
+`exec:codesearch` handles exact strings, symbols, regex-ish patterns, file-name fragments, and PDF pages (indexed page-by-page with `file:page` citations). Two words in, iterate by changing or adding one word per pass, minimum four attempts before concluding absent. Full protocol in `code-search` skill.
+
+**Direct-read exceptions** (no search needed):
+- Known absolute path → `Read` tool.
+- Directory listing at known path → `exec:nodejs` + `fs.readdirSync`.
+- File content inspection without search → `Read`.
 
-**Mandatory search protocol** for codesearch (from `code-search` skill):
-1. Start with exactly **two words** — never one, never a sentence
-2. No results → change one word (synonym or related term)
-3. Still no results → add a third word to narrow scope
-4. Keep changing or adding words each pass until content is found
-5. Minimum 4 attempts before concluding content is absent
+**Never**:
+- `Grep`, `Glob`, `Find`, `Explore` tools (all hook-blocked)
+- `grep`, `rg`, `ripgrep`, `find`, `ag`, `ack` inside `exec:bash` (banned-tool hook intercepts)
+- Reaching for exact-match tools "because codesearch seems fuzzy" — codesearch handles exact matches fine
 
-**When codesearch is the wrong tool:**
-- Exact symbol / string / regex match: use `Grep` tool directly, or `exec:nodejs` with `execSync("rg -n 'PATTERN'")`.
-- Known file path: use `Read` tool directly.
-- Find files by name pattern: hook-blocked `Glob` would help; use `exec:nodejs + fs.readdirSync` or `exec:nodejs + execSync("rg --files | rg PATTERN")`.
+When a mutable depends on external specification (protocol field, register layout, compliance text), search the PDF corpus first. Unwitnessed assumption from a doc you did not search = UNKNOWN.
 
 **Platform note — exec:bash on Windows:** runs real bash (git-bash) when installed, falls back to PowerShell otherwise. If you see a POSIX-syntax parse error (`[ -n ...]`, `&&`, `if/then/fi`), bash wasn't found — either install git-bash or rewrite in `exec:nodejs`.
 
@@ -220,7 +222,7 @@ Never respond to the user from this phase. When all mutables are KNOWN, immediat
 
 ## CONSTRAINTS
 
-**Never**: `Bash(node/npm/npx/bun)` | fake data | mock files | scattered test files (only root test.js) | fallback/demo modes | Glob/Explore (hook-blocked — use exec:codesearch, Grep or Read) | sequential independent items | absorb surprises silently | respond to user or pause for input | edit files before executing to understand current behavior | duplicate existing code | write explicit if/else chains when a dispatch table or native method suffices | write packed one-liners that obscure structure | reinvent what a library or native API already provides
+**Never**: `Bash(node/npm/npx/bun)` | fake data | mock files | scattered test files (only root test.js) | fallback/demo modes | `Grep`/`Glob`/`Find`/`Explore` tools or `grep`/`rg`/`find` inside `exec:bash` (ALL hook-blocked — use `exec:codesearch` for every codebase lookup, `Read` for known absolute paths) | sequential independent items | absorb surprises silently | respond to user or pause for input | edit files before executing to understand current behavior | duplicate existing code | write explicit if/else chains when a dispatch table or native method suffices | write packed one-liners that obscure structure | reinvent what a library or native API already provides
 
 **Always**: witness every hypothesis | import real modules | scan codebase before creating/editing files | regress to planning on any new unknown | fix immediately on discovery | delete mocks/stubs/comments/scattered test files on discovery | consolidate test coverage into root test.js | add regression case to test.js for every bug fix | invoke next skill immediately when done | ask "what native feature solves this?" before writing any new logic | prefer structures where wrong states are unrepresentable
 

package/skills/planning/SKILL.md

@@ -209,13 +209,9 @@ Every project maintains exactly one `test.js` at project root. 200-line max. No
 
 ## RESPONSE POLICY
 
-Terse like smart caveman. Technical substance stays. Fluff dies. Default: **full**. Switch: `/caveman lite|full|ultra`.
+Always terse. Technical substance stays. Fluff dies. Drop articles, filler, pleasantries, hedging. Fragments OK. Short synonyms. Technical terms exact. Pattern: `[thing] [action] [reason]. [next step].`
 
-Drop: articles, filler, pleasantries, hedging. Fragments OK. Short synonyms. Technical terms exact. Code unchanged. Pattern: `[thing] [action] [reason]. [next step].`
-
-Levels: **lite** = no filler, full sentences | **full** = drop articles, fragments OK | **ultra** = abbreviate all, arrows for causality | **wenyan-full** = 文言文, 80-90% compression | **wenyan-ultra** = max classical terse.
-
-Auto-Clarity: drop caveman for security warnings, irreversible confirmations, ambiguous sequences. Resume after. Code/commits/PRs write normal. "stop caveman" / "normal mode": revert.
+Code, commits, and PR descriptions write in normal prose. Security warnings, destructive confirmations, and genuinely ambiguous sequences also drop terseness. Everything else stays terse.
 
 ## CONSTRAINTS