[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74ZKZRFD3884418SYF0HFS6":3,"skill-same:01M74ZKZRFD3884418SYF0HFS6":58},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":31,"commitSha":32,"skillDirectory":33,"forksCount":34,"repositoryHtmlUrl":35,"repositoryHomepageUrl":36,"ownerHtmlUrl":37,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":38,"install":39,"resources":51},"01M74ZKZRFD3884418SYF0HFS6","public-surface-audit",null,"Scans git-tracked (public) files for operator-private tokens that should live only in gitignored files — real usernames, absolute home paths, companion-store names, company asset names. Reports file:line + matched token + severity, so a public\u002Fprivate split stays clean before publish. Triggered by \"public surface audit\", \"did I leak anything\", \"check tracked files for private tokens\", \"private token scan\", \"public-surface-audit\".","chrono-meta\u002Fforge-harness","chrono-meta","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F151837060?v=4","01M74ZDXK87ES4B30NKYSEH3MS","01M74ZDXM0F1NMYK7BS7PYR54G",33,"MIT","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fplugins\u002Ffh-meta\u002Fskills\u002Fpublic-surface-audit\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: public-surface-audit\ndescription: >-\n  Scans git-tracked (public) files for operator-private tokens that should live only in gitignored files — real usernames, absolute home paths, companion-store names, company asset names. Reports file:line + matched token + severity, so a public\u002Fprivate split stays clean before publish. Triggered by \"public surface audit\", \"did I leak anything\", \"check tracked files for private tokens\", \"private token scan\", \"public-surface-audit\".\nuser-invocable: true\nallowed-tools: [\"Read\", \"Bash\", \"Grep\", \"Glob\"]\nmodel: sonnet\ncategory: Composability Gate\n---\n\n# public-surface-audit — Operator-Private Token Leak Scan\n\nScans the git-tracked file set (the public surface) for operator-private tokens that were supposed\nto stay in gitignored files (e.g. `CLAUDE.local.md`, companion store). After a public\u002Fprivate split,\na front-door fix is not enough — a leaked username or absolute home path anywhere in the tracked set\nbreaks the \"public repo = model-agnostic methodology only\" invariant.\n\n> While `marketplace-gate` Check 5 answers \"is this repo broadly safe to publish?\" (API keys, internal\n> domains, license), `public-surface-audit` answers a narrower question: \"did any operator-private\n> token survive the public\u002Fprivate split into a tracked file?\" It scans `git ls-files` only — gitignored\n> files like `CLAUDE.local.md` are intentionally out of scope (they are the *correct* home for these tokens).\n\n## Triggers\n\n- `\u002Fpublic-surface-audit`\n- `\u002Fpublic-surface-audit --target \u003Crepo path>`\n- `\u002Fpublic-surface-audit --json` (machine-parseable verdict for hook-gating — see Step 5)\n- \"Did I leak anything into the public repo?\", \"public surface audit\", \"private token scan\"\n- \"Check tracked files for private tokens\", \"is my public\u002Fprivate split clean?\"\n- \"Did any operator-private token survive into a tracked file?\", \"scan before publish\"\n\n---\n\n## Scope — Tracked Files Only\n\nThis skill scans **only `git ls-files`** (committed\u002Fstaged tracked files). Gitignored files are\ndeliberately excluded — `CLAUDE.local.md`, the companion store, and local session data are the\n*correct* home for operator-private tokens, so finding them there is not a leak.\n\n```bash\nARGS=\"${ARGUMENTS\u002F\u002F--json\u002F}\"            # strip flags first — \"--json\" alone must not become the path\nREPO_PATH=\"${ARGS#--target }\"\nREPO_PATH=\"$(printf '%s' \"$REPO_PATH\" | xargs)\"\nREPO_PATH=\"${REPO_PATH:-$(pwd)}\"\ngit -C \"$REPO_PATH\" rev-parse --is-inside-work-tree >\u002Fdev\u002Fnull 2>&1 \\\n  || { echo \"Not a git repo — public-surface-audit scans git-tracked files only. Aborting.\"; exit 1; }\necho \"Target: $REPO_PATH\"\ngit -C \"$REPO_PATH\" ls-files | wc -l | xargs echo \"Tracked files:\"\n```\n\n---\n\n## Step 1. Pattern List (configurable)\n\nThe pattern source is **two layers, matching the mechanical scan layer** (`scripts\u002Fpsa_scan_lib.sh`,\nshared with the pre-commit\u002Fpre-push\u002Fpre-publish hooks — this skill must render the same verdict the\nhooks would, or the same repo state gets two answers):\n\n1. **`.claude\u002Frules\u002F.public-surface-patterns.defaults`** — committed, universal placeholder-grade\n   patterns (safe to publish; carries no operator literal).\n2. **`.claude\u002Frules\u002F.public-surface-patterns`** — **gitignored, operator literals** (or `PSA_PATTERNS`\n   env override). Your real username and employer name live *only* here — hardcoding them in a tracked\n   file would leak exactly what this skill hunts. One `severity\u003CTAB>regex` per line.\n\nThe scan reads the **union of both layers**. This SKILL.md carries only placeholders; the skill\ndogfoods its own rule.\n\n| # | Token class | Severity | Placeholder (real value goes in the gitignored source) | Why private |\n|:-:|---|:-:|---|---|\n| 1 | Real username | HIGH | `\u003Cyour-unix-username>` | Personal identity — must not appear on public surface |\n| 2 | Company \u002F employer asset name | HIGH | `\u003Ccompany-asset>` (alternation OK) | Company-confidential, leak-forbidden |\n| 3 | Operator absolute home path | MED | `\u002FUsers\u002F[a-z0-9_-]+\u002F` (generic — carries no literal name) | Machine-bound, leaks username + local layout |\n| 4 | Companion-store name | LOW | `\u003Ccompanion-store-name>` | Private companion store — methodology should not name it |\n| 5 | Operator-private script \u002F handoff name | LOW | `\u003Cprivate-script>`, `\u003Cprivate-dir>\u002F` | Operator-specific wiring, belongs in `CLAUDE.local.md` |\n\n**Severity meaning**:\n- **HIGH** — real name or company asset. A leak is a confidentiality \u002F identity exposure. Block publish.\n- **MED** — absolute home path. Leaks username + local filesystem layout; also a portability bug.\n- **LOW** — companion-store \u002F private-wiring name. Methodology should be model-agnostic; naming a private\n  store is drift, not a confidentiality breach.\n\n> **Setup**: put your real values in the gitignored override (one `severity\u003CTAB>regex` per line).\n> Degrade states, by layer — never collapse them:\n> - **Neither layer present** → **NOT CONFIGURED** — *not* CLEAN. A missing pattern source must never\n>   masquerade as a clean bill of health (\"nothing scanned\" misread as \"nothing leaked\").\n> - **Defaults only** (override absent) → scan runs, but the verdict carries the label\n>   `operator literals NOT CONFIGURED — universal patterns only; identity\u002Fcompany classes UNSCANNED`.\n>   This mirrors what the hooks scan in the same state, instead of reporting NOT CONFIGURED while the\n>   hook path quietly scans defaults (two verdicts for one repo state — the divergence this line closes).\n> - To declare \"I genuinely have no private tokens\", create the override **empty** — an empty file is\n>   an explicit declaration, an absent file is unconfigured.\n\n---\n\n## Step 2. Allowlist (legitimate references)\n\nSome tracked files legitimately reference otherwise-private tokens — the scan must not flag these as\nleaks. Maintain an allowlist of `file path :: token` pairs. A match is suppressed only when **both**\nthe file and the token are on the allowlist row.\n\n**Where the filled allowlist lives**: a populated allowlist row *names private tokens*, so it belongs\nin a **gitignored** source beside the pattern override (`.claude\u002Frules\u002F.public-surface-allowlist`, or a\nsection of `CLAUDE.local.md`) — never in a tracked file. The table below shows generic shapes only; a\ntracked copy of your real allowlist would itself be the leak this skill scans for.\n\n| Tracked file | Allowed tokens | Reason |\n|---|---|---|\n| `.gitignore` | companion-store name, sync-script name | Must name what it ignores |\n| your sync script (e.g. `scripts\u002F\u003Csync-script>`) | companion-store name, operator-dir names, home path | The sync script's whole job is the companion store |\n| an install-template rules file | home path, companion-store name | Install template — the install path is its content |\n| a doc describing the companion-store *pattern* | the `*-be` pattern (no literal store name) | Documents the pattern generically |\n\n**Allowlist rule**: a hit on file F matching token T is **suppressed** iff a row exists with file == F\nand T in that row's allowed tokens. Everything else is reported. Keep the allowlist tight — when in\ndoubt, report and let the user confirm. Genuinely model-agnostic mentions (the `*-be` companion\n*pattern* without the literal store name) should not require allowlisting because they do not match a\nliteral private token.\n\n---\n\n## Step 3. Scan\n\nFor each pattern in Step 1, grep the tracked set, then drop allowlisted hits. Do this for **every**\npattern row before producing the report — do not stop at the first HIT.\n\n**Binary \u002F generated carve-out**: skip binaries. Additionally note (do not auto-suppress) hits inside\ngenerated artifacts (e.g. `paper\u002F*.html` exported from a private source) — these are real leaks on the\npublic surface and must be reported, but the fix is \"regenerate from a sanitized source,\" not \"edit the\nHTML by hand.\" Flag them with a `(generated artifact)` note.\n\n> **Detail**: See `SKILL_detail.md §Step3-Scan-Script` — the grep-per-pattern bash implementation,\n> flag reference (`-n`\u002F`-I`\u002F`-E`) — read when executing this step.\n\n---\n\n## Step 3b. FP Hygiene — Placeholder & Example Exclusion\n\nA scan that flags its own placeholders erodes trust. Two **value-shape** classes are never real leaks\nand are dropped before the report — imported from `gstack-redact`'s canonical-example allowlist, scoped\nin PSA's direction to the *matched token* (not the whole line, so a real leak sharing a substring still\nreports):\n\n- **Angle-bracket placeholders** — the matched token is itself a placeholder (`\u003Cyour-unix-username>`,\n  `\u003Ccompany-asset>`, `{project}`). PSA dogfoods these in Step 1; the scan must not report them as leaks.\n- **Canonical dummy values** — the matched token is a documented example\u002Fdummy (`EXAMPLE`, `dummy`,\n  `changeme`, `REDACTED`, `xxxx`, AWS-doc keys like `AKIAIOSFODNN7EXAMPLE`). A high-entropy *example* is\n  not a secret.\n\nThis differs from the Step 2 allowlist: Step 2 suppresses by **file::token legitimacy**, Step 3b by\n**token value-shape**. Both run — Step 2 then Step 3b. Keep it tight (PSA's \"allowlist tight\" rule): if a\ntoken only *contains* an example substring but is otherwise a real private value, it still reports.\n\n> **Detail**: See `SKILL_detail.md §Step3b-FP-Hygiene-Script` — the token-shape-only extraction bash\n> implementation (never line-level) — read when executing this step.\n\n---\n\n## Step 3c. Ignore-Verification — the gitignore-mistake safeguard (root-cause, runs before Step 4)\n\nSteps 3\u002F3b scan what **is** tracked. This step checks the inverse and the upstream cause: a file you\n*intend* to be private that is **not actually ignored** — the gitignore mistake that becomes tomorrow's\ntracked leak. For each expected-private path, `git check-ignore -v` reports whether it is ignored **and\nby which layer** (committed `.gitignore` · local `.git\u002Finfo\u002Fexclude` · global `core.excludesFile`). A\npath that resolves to no ignore rule is a MISS — surface it before it is ever `git add`ed.\n\n- **OK** — ignored; the report names which layer covers it (operator-private files should resolve to\n  `.git\u002Finfo\u002Fexclude` or `.gitignore`, never end up tracked).\n- **MISS** — exists but no ignore rule covers it → counts as **🟠 MED** in the Step 4 verdict (a latent\n  leak, not yet a breach).\n- **TRACKED** — already committed: Step 3 scans its *contents*; this also flags it for un-tracking if it\n  was meant to be private. Tracked status is tested **before** ignore status — a file can match an\n  ignore rule yet still be force-added (`git add -f`), the exact ignored-but-committed mechanism behind\n  the PR #109 leak, so TRACKED must win over a false-clean OK.\n\nWhy this is the safeguard for **gitignore mistakes** (a wrong assumption about what is ignored):\n`.gitignore` is committed\u002Fshared, `.git\u002Finfo\u002Fexclude` is local\u002Fpersonal, and a global `core.excludesFile`\nignores across all repos — `git check-ignore -v` is the one command that says *which* rule (if any)\napplies, so an \"I thought it was ignored\" error surfaces here instead of in a public PR (the PR #109\nclass of leak). Diagnostic-only: this step never writes — it reports, the operator adds the ignore rule.\n\n> **Detail**: See `SKILL_detail.md §Step3c-Ignore-Verification-Script` — the expected-private-path\n> enumeration + `git check-ignore`\u002F`git ls-files --error-unmatch` bash implementation, bash\u002Fzsh\n> portability note — read when executing this step.\n\n---\n\n## Step 4. Report\n\nReport per-hit `file:line → matched token [class]` grouped by severity, then the overall verdict.\n**Verdict thresholds** (behavioral — these define the gate):\n\n- ⚪ **NOT CONFIGURED** — no pattern layer present (nothing scanned — NOT a clean result; set up first)\n- 🟢 **CLEAN** — pattern source present (incl. empty override), 0 HIGH + 0 MED + 0 LOW (after allowlist)\n- 🟡 **REVIEW** — 0 HIGH + 0 MED, LOW-only (drift, not a breach)\n- 🔴 **LEAK** — 1+ HIGH or 1+ MED (block publish \u002F fix before commit)\n\n**Every verdict line carries a `coverage:` qualifier** — `full` (both layers loaded) or\n`defaults-only` (override absent: operator identity\u002Fcompany classes UNSCANNED). The qualifier exists\nbecause the enum alone cannot preserve the defaults-only state: `CLEAN (coverage: defaults-only)` on\nan operator-configured checkout is missing evidence, not a pass — publish-gating callers there must\nrequire `coverage: full`.\n\n> **Detail**: See `SKILL_detail.md §Report-Template` — the full human-report ASCII layout (severity\n> buckets, allowlist-suppressed line) — read when formatting the Step 4 report.\n\nPer HIGH\u002FMED hit, append a one-line prescription:\n- **HIGH (username\u002Fcompany)** — move the line to `CLAUDE.local.md` (or regenerate the artifact from a\n  sanitized source); never edit-in-place if it is a generated file.\n- **MED (absolute path)** — replace with a relative path, a `~`-anchored path, or a `{project}`\n  placeholder; absolute home paths are also a portability bug for external clones.\n- **LOW (companion-store\u002Fwiring)** — rephrase to the model-agnostic *pattern* (e.g. \"a private companion\n  store\" \u002F \"the `*-be` companion pattern\") instead of the literal name, unless the file is the\n  `.gitignore` \u002F sync script that must name it (allowlist those).\n\n**Simplification guard**: 🟢 CLEAN → collapse the report to one line: \"public surface clean — 0 private\ntokens in {N} tracked files (X allowlist-suppressed).\" Do not print empty severity buckets.\n\n---\n\n## Step 5. Machine Output (`--json`) — Hook-Gateable Verdict\n\nBy default PSA prints the Step 4 human report. With `--json`, emit a machine-parseable verdict so a\n**pre-publish \u002F pre-push hook can gate on counts mechanically** — turning PSA from advisory into\nenforceable (FH's \"enforcement is a hook, not a prompt\" principle). Imported from `gstack-redact --json`.\n\n> **Detail**: See `SKILL_detail.md §JSON-Schema` — the `--json` output object shape (findings array,\n> counts, verdict field) — read when emitting or parsing the machine verdict.\n\n`verdict` is one of `CLEAN | REVIEW | LEAK | NOT_CONFIGURED` (same thresholds as Step 4). **`verdict` is\nauthoritative — never gate on `counts` alone**: a counts-only check (`HIGH==0 && MED==0`) misreads\n`NOT_CONFIGURED` (which also has zero counts) as a pass. A caller blocks when `verdict` is `LEAK` **or**\n`NOT_CONFIGURED` — an unconfigured scan is not a pass (the same silent-failure guard as the human path:\nabsence ≠ CLEAN). The JSON carries the same `coverage` field as Step 4; on an operator-configured\ncheckout a publish-gating caller additionally blocks on `coverage: \"defaults-only\"` (identity\u002Fcompany\nclasses unscanned there — zero counts prove nothing about the classes that never ran).\n\n---\n\n## Connected Skills\n\n| Situation | Connected skill |\n|---|---|\n| Broader pre-publish repo readiness (README, license, API keys) | `\u002Fmarketplace-gate` (Check 5 Public Safety is the wide net; this skill is the private-token detail) |\n| A leak is a recurring process gap, not a one-off | log via `field-harvest` → candidate `#rule-candidate` |\n| Where should the leaked content actually live? | `\u002Fasset-placement-gate` (hub vs project vs CLAUDE.local.md) |\n| Phantom refs \u002F stale links on the same surface | `\u002Fphantom-quench` (forward axis — orthogonal to this leak axis) |\n\n---\n\n## External User Environment Adaptation\n\nUsable standalone — no hub clone required.\n- **No companion store \u002F no operator-private tokens** → create the pattern source **empty** to declare\n  this explicitly; the scan then reports CLEAN. Leaving the file *absent* instead yields NOT CONFIGURED\n  (a deliberate distinction — absence is \"unknown\", not \"clean\"). The skill is only useful once you have\n  a public\u002Fprivate split to protect.\n- **No `.gitignore` allowlist needs** → Step 2 allowlist may be empty; every hit is then reported.\n\n---\n\n## Done When\n\n```\nStep 1 pattern source resolved (both layers checked; absent\u002Fpartial\n  states rendered as their labeled degrade state, never as CLEAN)      — mandatory-pass\n+ Step 2 allowlist applied (from a gitignored source)                  — mandatory-pass\n+ Step 3 scan run for every pattern over git ls-files (tracked only —\n  gitignored excluded), malformed pattern rows counted and reported,\n  never silently dropped                                               — measured (hit + malformed counts)\n+ Step 3b\u002F3c run (FP-hygiene on matched tokens; ignore-verification\n  for expected-private paths) or explicitly skipped with reason        — mandatory-pass\n+ Step 4 report output: per-hit file:line + token + severity, plus\n  overall verdict                                                      — mandatory-pass\n+ \"public-surface-audit Complete\" declaration output                   — mandatory-pass\n```\n\nVerdict: **CLEAN** (0 tokens after allowlist) | **REVIEW** (LOW-only — drift, prescriptions noted) |\n**LEAK** (1+ HIGH or 1+ MED — block publish, prescriptions attached) | **NOT_CONFIGURED** (no pattern\nsource — nothing scanned, not a pass; same 4-value enum as Step 5 `--json`. An earlier version of this\nenum had 3 values here, so a session following only Done When would force absent-config into CLEAN\u002F\nREVIEW\u002FLEAK — re-opening the exact silent-pass the body closes).\n\n---\n\n## Operating Notes\n\n- **Tracked-only is the point**: never scan gitignored files. A token in `CLAUDE.local.md` is correct\n  placement, not a leak — scanning it would produce false LEAK verdicts and erode trust in the skill.\n- **Patterns are data, not code**: the Step 1 table is the configurable surface. A user with a different\n  username\u002Femployer\u002Fcompanion-store edits the table; the scan logic is unchanged.\n- **Generated artifacts are real leaks**: an exported HTML\u002FPDF carrying a username is still a public-surface\n  leak even though hand-editing it is wrong — report it, prescribe \"regenerate from sanitized source\".\n- **Allowlist tight, not loose**: when unsure whether a reference is legitimate, report it. A false LEAK\n  the user dismisses is cheaper than a real leak suppressed by an over-broad allowlist.\n- **Auto-redact deliberately not imported**: `gstack-redact` offers `--auto-redact` (rewrite + diff). PSA's\n  philosophy is *report + prescribe; the human decides where the line goes* — auto-redacting a HIGH\n  (username\u002Fcompany) hit would pre-empt that judgment, and auto-editing a generated artifact is explicitly\n  wrong (regenerate from source). If ever imported, restrict to the MED absolute-home-path class only\n  (mechanically safe: `\u002FUsers\u002F\u003Cuser>\u002F` → `~\u002F` or `{project}`), never HIGH, never generated files.\n\n---\n\n## Sister-Asset Provenance\n\nStep 3b (FP hygiene) and Step 5 (`--json`) were imported from **garrytan\u002Fgstack** `gstack-redact` during\na 2026-06-06 sister-asset cross-audit.\n\n> **Detail**: See `SKILL_detail.md §Sister-Asset-Provenance` — the full import provenance (source file,\n> ontology adaptation, orthogonal-coverage rationale, reverse-contribution candidate) — read when tracing\n> or extending the gstack-redact lineage.\n","# public-surface-audit — Operator-Private Token Leak Scan\n\nScans the git-tracked file set (the public surface) for operator-private tokens that were supposed\nto stay in gitignored files (e.g. `CLAUDE.local.md`, companion store). After a public\u002Fprivate split,\na front-door fix is not enough — a leaked username or absolute home path anywhere in the tracked set\nbreaks the \"public repo = model-agnostic methodology only\" invariant.\n\n> While `marketplace-gate` Check 5 answers \"is this repo broadly safe to publish?\" (API keys, internal\n> domains, license), `public-surface-audit` answers a narrower question: \"did any operator-private\n> token survive the public\u002Fprivate split into a tracked file?\" It scans `git ls-files` only — gitignored\n> files like `CLAUDE.local.md` are intentionally out of scope (they are the *correct* home for these tokens).\n\n## Triggers\n\n- `\u002Fpublic-surface-audit`\n- `\u002Fpublic-surface-audit --target \u003Crepo path>`\n- `\u002Fpublic-surface-audit --json` (machine-parseable verdict for hook-gating — see Step 5)\n- \"Did I leak anything into the public repo?\", \"public surface audit\", \"private token scan\"\n- \"Check tracked files for private tokens\", \"is my public\u002Fprivate split clean?\"\n- \"Did any operator-private token survive into a tracked file?\", \"scan before publish\"\n\n---\n\n## Scope — Tracked Files Only\n\nThis skill scans **only `git ls-files`** (committed\u002Fstaged tracked files). Gitignored files are\ndeliberately excluded — `CLAUDE.local.md`, the companion store, and local session data are the\n*correct* home for operator-private tokens, so finding them there is not a leak.\n\n```bash\nARGS=\"${ARGUMENTS\u002F\u002F--json\u002F}\"            # strip flags first — \"--json\" alone must not become the path\nREPO_PATH=\"${ARGS#--target }\"\nREPO_PATH=\"$(printf '%s' \"$REPO_PATH\" | xargs)\"\nREPO_PATH=\"${REPO_PATH:-$(pwd)}\"\ngit -C \"$REPO_PATH\" rev-parse --is-inside-work-tree >\u002Fdev\u002Fnull 2>&1 \\\n  || { echo \"Not a git repo — public-surface-audit scans git-tracked files only. Aborting.\"; exit 1; }\necho \"Target: $REPO_PATH\"\ngit -C \"$REPO_PATH\" ls-files | wc -l | xargs echo \"Tracked files:\"\n```\n\n---\n\n## Step 1. Pattern List (configurable)\n\nThe pattern source is **two layers, matching the mechanical scan layer** (`scripts\u002Fpsa_scan_lib.sh`,\nshared with the pre-commit\u002Fpre-push\u002Fpre-publish hooks — this skill must render the same verdict the\nhooks would, or the same repo state gets two answers):\n\n1. **`.claude\u002Frules\u002F.public-surface-patterns.defaults`** — committed, universal placeholder-grade\n   patterns (safe to publish; carries no operator literal).\n2. **`.claude\u002Frules\u002F.public-surface-patterns`** — **gitignored, operator literals** (or `PSA_PATTERNS`\n   env override). Your real username and employer name live *only* here — hardcoding them in a tracked\n   file would leak exactly what this skill hunts. One `severity\u003CTAB>regex` per line.\n\nThe scan reads the **union of both layers**. This SKILL.md carries only placeholders; the skill\ndogfoods its own rule.\n\n| # | Token class | Severity | Placeholder (real value goes in the gitignored source) | Why private |\n|:-:|---|:-:|---|---|\n| 1 | Real username | HIGH | `\u003Cyour-unix-username>` | Personal identity — must not appear on public surface |\n| 2 | Company \u002F employer asset name | HIGH | `\u003Ccompany-asset>` (alternation OK) | Company-confidential, leak-forbidden |\n| 3 | Operator absolute home path | MED | `\u002FUsers\u002F[a-z0-9_-]+\u002F` (generic — carries no literal name) | Machine-bound, leaks username + local layout |\n| 4 | Companion-store name | LOW | `\u003Ccompanion-store-name>` | Private companion store — methodology should not name it |\n| 5 | Operator-private script \u002F handoff name | LOW | `\u003Cprivate-script>`, `\u003Cprivate-dir>\u002F` | Operator-specific wiring, belongs in `CLAUDE.local.md` |\n\n**Severity meaning**:\n- **HIGH** — real name or company asset. A leak is a confidentiality \u002F identity exposure. Block publish.\n- **MED** — absolute home path. Leaks username + local filesystem layout; also a portability bug.\n- **LOW** — companion-store \u002F private-wiring name. Methodology should be model-agnostic; naming a private\n  store is drift, not a confidentiality breach.\n\n> **Setup**: put your real values in the gitignored override (one `severity\u003CTAB>regex` per line).\n> Degrade states, by layer — never collapse them:\n> - **Neither layer present** → **NOT CONFIGURED** — *not* CLEAN. A missing pattern source must never\n>   masquerade as a clean bill of health (\"nothing scanned\" misread as \"nothing leaked\").\n> - **Defaults only** (override absent) → scan runs, but the verdict carries the label\n>   `operator literals NOT CONFIGURED — universal patterns only; identity\u002Fcompany classes UNSCANNED`.\n>   This mirrors what the hooks scan in the same state, instead of reporting NOT CONFIGURED while the\n>   hook path quietly scans defaults (two verdicts for one repo state — the divergence this line closes).\n> - To declare \"I genuinely have no private tokens\", create the override **empty** — an empty file is\n>   an explicit declaration, an absent file is unconfigured.\n\n---\n\n## Step 2. Allowlist (legitimate references)\n\nSome tracked files legitimately reference otherwise-private tokens — the scan must not flag these as\nleaks. Maintain an allowlist of `file path :: token` pairs. A match is suppressed only when **both**\nthe file and the token are on the allowlist row.\n\n**Where the filled allowlist lives**: a populated allowlist row *names private tokens*, so it belongs\nin a **gitignored** source beside the pattern override (`.claude\u002Frules\u002F.public-surface-allowlist`, or a\nsection of `CLAUDE.local.md`) — never in a tracked file. The table below shows generic shapes only; a\ntracked copy of your real allowlist would itself be the leak this skill scans for.\n\n| Tracked file | Allowed tokens | Reason |\n|---|---|---|\n| `.gitignore` | companion-store name, sync-script name | Must name what it ignores |\n| your sync script (e.g. `scripts\u002F\u003Csync-script>`) | companion-store name, operator-dir names, home path | The sync script's whole job is the companion store |\n| an install-template rules file | home path, companion-store name | Install template — the install path is its content |\n| a doc describing the companion-store *pattern* | the `*-be` pattern (no literal store name) | Documents the pattern generically |\n\n**Allowlist rule**: a hit on file F matching token T is **suppressed** iff a row exists with file == F\nand T in that row's allowed tokens. Everything else is reported. Keep the allowlist tight — when in\ndoubt, report and let the user confirm. Genuinely model-agnostic mentions (the `*-be` companion\n*pattern* without the literal store name) should not require allowlisting because they do not match a\nliteral private token.\n\n---\n\n## Step 3. Scan\n\nFor each pattern in Step 1, grep the tracked set, then drop allowlisted hits. Do this for **every**\npattern row before producing the report — do not stop at the first HIT.\n\n**Binary \u002F generated carve-out**: skip binaries. Additionally note (do not auto-suppress) hits inside\ngenerated artifacts (e.g. `paper\u002F*.html` exported from a private source) — these are real leaks on the\npublic surface and must be reported, but the fix is \"regenerate from a sanitized source,\" not \"edit the\nHTML by hand.\" Flag them with a `(generated artifact)` note.\n\n> **Detail**: See `SKILL_detail.md §Step3-Scan-Script` — the grep-per-pattern bash implementation,\n> flag reference (`-n`\u002F`-I`\u002F`-E`) — read when executing this step.\n\n---\n\n## Step 3b. FP Hygiene — Placeholder & Example Exclusion\n\nA scan that flags its own placeholders erodes trust. Two **value-shape** classes are never real leaks\nand are dropped before the report — imported from `gstack-redact`'s canonical-example allowlist, scoped\nin PSA's direction to the *matched token* (not the whole line, so a real leak sharing a substring still\nreports):\n\n- **Angle-bracket placeholders** — the matched token is itself a placeholder (`\u003Cyour-unix-username>`,\n  `\u003Ccompany-asset>`, `{project}`). PSA dogfoods these in Step 1; the scan must not report them as leaks.\n- **Canonical dummy values** — the matched token is a documented example\u002Fdummy (`EXAMPLE`, `dummy`,\n  `changeme`, `REDACTED`, `xxxx`, AWS-doc keys like `AKIAIOSFODNN7EXAMPLE`). A high-entropy *example* is\n  not a secret.\n\nThis differs from the Step 2 allowlist: Step 2 suppresses by **file::token legitimacy**, Step 3b by\n**token value-shape**. Both run — Step 2 then Step 3b. Keep it tight (PSA's \"allowlist tight\" rule): if a\ntoken only *contains* an example substring but is otherwise a real private value, it still reports.\n\n> **Detail**: See `SKILL_detail.md §Step3b-FP-Hygiene-Script` — the token-shape-only extraction bash\n> implementation (never line-level) — read when executing this step.\n\n---\n\n## Step 3c. Ignore-Verification — the gitignore-mistake safeguard (root-cause, runs before Step 4)\n\nSteps 3\u002F3b scan what **is** tracked. This step checks the inverse and the upstream cause: a file you\n*intend* to be private that is **not actually ignored** — the gitignore mistake that becomes tomorrow's\ntracked leak. For each expected-private path, `git check-ignore -v` reports whether it is ignored **and\nby which layer** (committed `.gitignore` · local `.git\u002Finfo\u002Fexclude` · global `core.excludesFile`). A\npath that resolves to no ignore rule is a MISS — surface it before it is ever `git add`ed.\n\n- **OK** — ignored; the report names which layer covers it (operator-private files should resolve to\n  `.git\u002Finfo\u002Fexclude` or `.gitignore`, never end up tracked).\n- **MISS** — exists but no ignore rule covers it → counts as **🟠 MED** in the Step 4 verdict (a latent\n  leak, not yet a breach).\n- **TRACKED** — already committed: Step 3 scans its *contents*; this also flags it for un-tracking if it\n  was meant to be private. Tracked status is tested **before** ignore status — a file can match an\n  ignore rule yet still be force-added (`git add -f`), the exact ignored-but-committed mechanism behind\n  the PR #109 leak, so TRACKED must win over a false-clean OK.\n\nWhy this is the safeguard for **gitignore mistakes** (a wrong assumption about what is ignored):\n`.gitignore` is committed\u002Fshared, `.git\u002Finfo\u002Fexclude` is local\u002Fpersonal, and a global `core.excludesFile`\nignores across all repos — `git check-ignore -v` is the one command that says *which* rule (if any)\napplies, so an \"I thought it was ignored\" error surfaces here instead of in a public PR (the PR #109\nclass of leak). Diagnostic-only: this step never writes — it reports, the operator adds the ignore rule.\n\n> **Detail**: See `SKILL_detail.md §Step3c-Ignore-Verification-Script` — the expected-private-path\n> enumeration + `git check-ignore`\u002F`git ls-files --error-unmatch` bash implementation, bash\u002Fzsh\n> portability note — read when executing this step.\n\n---\n\n## Step 4. Report\n\nReport per-hit `file:line → matched token [class]` grouped by severity, then the overall verdict.\n**Verdict thresholds** (behavioral — these define the gate):\n\n- ⚪ **NOT CONFIGURED** — no pattern layer present (nothing scanned — NOT a clean result; set up first)\n- 🟢 **CLEAN** — pattern source present (incl. empty override), 0 HIGH + 0 MED + 0 LOW (after allowlist)\n- 🟡 **REVIEW** — 0 HIGH + 0 MED, LOW-only (drift, not a breach)\n- 🔴 **LEAK** — 1+ HIGH or 1+ MED (block publish \u002F fix before commit)\n\n**Every verdict line carries a `coverage:` qualifier** — `full` (both layers loaded) or\n`defaults-only` (override absent: operator identity\u002Fcompany classes UNSCANNED). The qualifier exists\nbecause the enum alone cannot preserve the defaults-only state: `CLEAN (coverage: defaults-only)` on\nan operator-configured checkout is missing evidence, not a pass — publish-gating callers there must\nrequire `coverage: full`.\n\n> **Detail**: See `SKILL_detail.md §Report-Template` — the full human-report ASCII layout (severity\n> buckets, allowlist-suppressed line) — read when formatting the Step 4 report.\n\nPer HIGH\u002FMED hit, append a one-line prescription:\n- **HIGH (username\u002Fcompany)** — move the line to `CLAUDE.local.md` (or regenerate the artifact from a\n  sanitized source); never edit-in-place if it is a generated file.\n- **MED (absolute path)** — replace with a relative path, a `~`-anchored path, or a `{project}`\n  placeholder; absolute home paths are also a portability bug for external clones.\n- **LOW (companion-store\u002Fwiring)** — rephrase to the model-agnostic *pattern* (e.g. \"a private companion\n  store\" \u002F \"the `*-be` companion pattern\") instead of the literal name, unless the file is the\n  `.gitignore` \u002F sync script that must name it (allowlist those).\n\n**Simplification guard**: 🟢 CLEAN → collapse the report to one line: \"public surface clean — 0 private\ntokens in {N} tracked files (X allowlist-suppressed).\" Do not print empty severity buckets.\n\n---\n\n## Step 5. Machine Output (`--json`) — Hook-Gateable Verdict\n\nBy default PSA prints the Step 4 human report. With `--json`, emit a machine-parseable verdict so a\n**pre-publish \u002F pre-push hook can gate on counts mechanically** — turning PSA from advisory into\nenforceable (FH's \"enforcement is a hook, not a prompt\" principle). Imported from `gstack-redact --json`.\n\n> **Detail**: See `SKILL_detail.md §JSON-Schema` — the `--json` output object shape (findings array,\n> counts, verdict field) — read when emitting or parsing the machine verdict.\n\n`verdict` is one of `CLEAN | REVIEW | LEAK | NOT_CONFIGURED` (same thresholds as Step 4). **`verdict` is\nauthoritative — never gate on `counts` alone**: a counts-only check (`HIGH==0 && MED==0`) misreads\n`NOT_CONFIGURED` (which also has zero counts) as a pass. A caller blocks when `verdict` is `LEAK` **or**\n`NOT_CONFIGURED` — an unconfigured scan is not a pass (the same silent-failure guard as the human path:\nabsence ≠ CLEAN). The JSON carries the same `coverage` field as Step 4; on an operator-configured\ncheckout a publish-gating caller additionally blocks on `coverage: \"defaults-only\"` (identity\u002Fcompany\nclasses unscanned there — zero counts prove nothing about the classes that never ran).\n\n---\n\n## Connected Skills\n\n| Situation | Connected skill |\n|---|---|\n| Broader pre-publish repo readiness (README, license, API keys) | `\u002Fmarketplace-gate` (Check 5 Public Safety is the wide net; this skill is the private-token detail) |\n| A leak is a recurring process gap, not a one-off | log via `field-harvest` → candidate `#rule-candidate` |\n| Where should the leaked content actually live? | `\u002Fasset-placement-gate` (hub vs project vs CLAUDE.local.md) |\n| Phantom refs \u002F stale links on the same surface | `\u002Fphantom-quench` (forward axis — orthogonal to this leak axis) |\n\n---\n\n## External User Environment Adaptation\n\nUsable standalone — no hub clone required.\n- **No companion store \u002F no operator-private tokens** → create the pattern source **empty** to declare\n  this explicitly; the scan then reports CLEAN. Leaving the file *absent* instead yields NOT CONFIGURED\n  (a deliberate distinction — absence is \"unknown\", not \"clean\"). The skill is only useful once you have\n  a public\u002Fprivate split to protect.\n- **No `.gitignore` allowlist needs** → Step 2 allowlist may be empty; every hit is then reported.\n\n---\n\n## Done When\n\n```\nStep 1 pattern source resolved (both layers checked; absent\u002Fpartial\n  states rendered as their labeled degrade state, never as CLEAN)      — mandatory-pass\n+ Step 2 allowlist applied (from a gitignored source)                  — mandatory-pass\n+ Step 3 scan run for every pattern over git ls-files (tracked only —\n  gitignored excluded), malformed pattern rows counted and reported,\n  never silently dropped                                               — measured (hit + malformed counts)\n+ Step 3b\u002F3c run (FP-hygiene on matched tokens; ignore-verification\n  for expected-private paths) or explicitly skipped with reason        — mandatory-pass\n+ Step 4 report output: per-hit file:line + token + severity, plus\n  overall verdict                                                      — mandatory-pass\n+ \"public-surface-audit Complete\" declaration output                   — mandatory-pass\n```\n\nVerdict: **CLEAN** (0 tokens after allowlist) | **REVIEW** (LOW-only — drift, prescriptions noted) |\n**LEAK** (1+ HIGH or 1+ MED — block publish, prescriptions attached) | **NOT_CONFIGURED** (no pattern\nsource — nothing scanned, not a pass; same 4-value enum as Step 5 `--json`. An earlier version of this\nenum had 3 values here, so a session following only Done When would force absent-config into CLEAN\u002F\nREVIEW\u002FLEAK — re-opening the exact silent-pass the body closes).\n\n---\n\n## Operating Notes\n\n- **Tracked-only is the point**: never scan gitignored files. A token in `CLAUDE.local.md` is correct\n  placement, not a leak — scanning it would produce false LEAK verdicts and erode trust in the skill.\n- **Patterns are data, not code**: the Step 1 table is the configurable surface. A user with a different\n  username\u002Femployer\u002Fcompanion-store edits the table; the scan logic is unchanged.\n- **Generated artifacts are real leaks**: an exported HTML\u002FPDF carrying a username is still a public-surface\n  leak even though hand-editing it is wrong — report it, prescribe \"regenerate from sanitized source\".\n- **Allowlist tight, not loose**: when unsure whether a reference is legitimate, report it. A false LEAK\n  the user dismisses is cheaper than a real leak suppressed by an over-broad allowlist.\n- **Auto-redact deliberately not imported**: `gstack-redact` offers `--auto-redact` (rewrite + diff). PSA's\n  philosophy is *report + prescribe; the human decides where the line goes* — auto-redacting a HIGH\n  (username\u002Fcompany) hit would pre-empt that judgment, and auto-editing a generated artifact is explicitly\n  wrong (regenerate from source). If ever imported, restrict to the MED absolute-home-path class only\n  (mechanically safe: `\u002FUsers\u002F\u003Cuser>\u002F` → `~\u002F` or `{project}`), never HIGH, never generated files.\n\n---\n\n## Sister-Asset Provenance\n\nStep 3b (FP hygiene) and Step 5 (`--json`) were imported from **garrytan\u002Fgstack** `gstack-redact` during\na 2026-06-06 sister-asset cross-audit.\n\n> **Detail**: See `SKILL_detail.md §Sister-Asset-Provenance` — the full import provenance (source file,\n> ontology adaptation, orthogonal-coverage rationale, reverse-contribution candidate) — read when tracing\n> or extending the gstack-redact lineage.\n",{"name":5,"model":23,"category":24,"description":7,"allowed-tools":25,"user-invocable":30},"sonnet","Composability Gate",[26,27,28,29],"Read","Bash","Grep","Glob",true,"main","899495d17a23f30ca149d4ac29a137a8fa7864e3","plugins\u002Ffh-meta\u002Fskills\u002Fpublic-surface-audit",1,"https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness","https:\u002F\u002Fchrono-meta.github.io\u002Fforge-harness\u002F","https:\u002F\u002Fgithub.com\u002Fchrono-meta","2026-11-15 02:42:21",{"source":8,"skillName":5,"commands":40,"prompt":50,"zipAvailable":30},[41,44,47],{"tool":42,"command":43},"npx","npx skills add chrono-meta\u002Fforge-harness --skill public-surface-audit",{"tool":45,"command":46},"bunx","bunx skills add chrono-meta\u002Fforge-harness --skill public-surface-audit",{"tool":48,"command":49},"pnpm","pnpm dlx skills add chrono-meta\u002Fforge-harness --skill public-surface-audit","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fplugins\u002Ffh-meta\u002Fskills\u002Fpublic-surface-audit\u002FSKILL.md",[52],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"SKILL_detail.md","REFERENCE","参考文档",11487,"https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fplugins\u002Ffh-meta\u002Fskills\u002Fpublic-surface-audit\u002FSKILL_detail.md",[59,66,73,80,86,93],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":18},"01M74ZPW36H1HVE0R61QZ5ZND0","todo-orchestrator","TODO CLI 프로젝트의 기능 추가·버그 수정·리팩터링·테스트 보강을 개발자와 QA 에이전트로 조율한다. todo 앱 관련 코드 작업 요청에 사용한다. 다시 실행, 재실행, 수정, 보완, 이전 결과 개선, QA만 다시 같은 후속 요청에도 사용한다. 단순 질문에는 사용하지 않는다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP4_r2\u002F.claude\u002Fskills\u002Ftodo-orchestrator\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":18},"01M74ZPS4EP2349B346R5JQEYE","todo-dev","TODO CLI(todo\u002Fcli.py)에 명령을 추가하거나 버그를 고치고 테스트를 작성하는 방법. todo 앱 코드 수정·기능 추가·테스트 작성·리팩터링 작업에 사용한다. 수정, 보완, 다시 구현 요청에도 사용한다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP4_r2\u002F.claude\u002Fskills\u002Ftodo-dev\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":18},"01M74ZPPX2743J6FPT8N2ZT0D5","todo-dev-team","TODO CLI(todo\u002Fcli.py) 개발 작업을 개발자·테스터·리뷰어 팀으로 진행하는 오케스트레이터. 명령 추가, 버그 수정, 테스트 작성, 리팩터링, 커밋·푸시 요청에 사용한다. 다시 실행, 재실행, 수정, 보완, 리뷰 반영, 테스트만 다시 같은 후속 요청에도 사용한다. 단순 질문에는 쓰지 않는다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP3_r2\u002F.claude\u002Fskills\u002Ftodo-dev-team\u002FSKILL.md",[],[],{"skillNo":81,"name":68,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":18},"01M74ZPKWST7DF24H0EN9RCEME","TODO CLI 개발 작업(명령·기능 추가, 버그 수정, 테스트 작성, 리뷰, 커밋·push)을 개발자·테스터·리뷰어 에이전트 팀으로 진행하는 오케스트레이터. todo CLI, todo.json, todo\u002Fcli.py 관련 구현·수정·테스트 요청에 사용한다. 재실행, 수정, 보완, 이전 결과 개선, 테스트만 다시, 리뷰만 다시 요청에도 사용한다. 단순 질문에는 쓰지 않는다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP3_r1\u002F.claude\u002Fskills\u002Ftodo-dev\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":18},"01M74ZPHBP62X3KKSV2RVKSBKT","weekly-summary-write","정리된 회의록으로 주간 회의 요약 문서를 작성하는 형식과 문체 규칙. 주간 요약 작성, 요약 문서 수정·보완, 검토 의견 반영 시 사용한다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP2_r2\u002F.claude\u002Fskills\u002Fweekly-summary-write\u002FSKILL.md",[],[],{"skillNo":94,"name":95,"nameZh":6,"description":96,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":18},"01M74ZPECSXY4C0QNHAD96Z88H","weekly-meeting-summary","주간 회의록을 정리하고 요약 문서를 만드는 전체 작업을 조율하는 오케스트레이터. 회의록 요약, 주간 회의 정리, 주간 보고용 요약 요청 시 사용한다. 다시 실행, 재실행, 요약 수정·보완, 이전 결과를 바탕으로 한 개선, 정리만\u002F검토만 다시 같은 후속 요청에도 사용한다.","https:\u002F\u002Fgithub.com\u002Fchrono-meta\u002Fforge-harness\u002Fblob\u002Fmain\u002Fscripts\u002Ffixtures\u002Fgenerated_harness\u002FP2_r2\u002F.claude\u002Fskills\u002Fweekly-meeting-summary\u002FSKILL.md",[],[]]