holocron/plugins/git/.apm/skills/pc-run/SKILL.md at 261e5b5491e971fe01bf64342482d531ecb7c916

Files

Defame1297 5e232503c4 feat(kyberforge): execute plugin-to-apm marketplace conversion

Why:
ADR-0015 established that Microsoft APM (apm.yml + .apm/) should replace
this repo's hand-authored plugin.json/marketplace.json model, with those
files becoming compiled output of `apm pack` instead of files edited by
hand via the (now-retired) plugin-author/marketplace-author skills.
Issue #90 was the deferred execution of that decision, gated on #88
(apm tooling) and #89 (apm-native agent-author/skill-author routing).

Implementation notes:
- All six plugins (bin, core, git, gitea, kyberforge, lint) now carry
  apm.yml + .apm/{skills,agents,hooks} as their authoring source. Skills
  moved with a plain git mv (content-identical across targets). Agents
  were re-authored, not moved: per ADR-0016, .apm/agents/*.agent.md
  compiles verbatim to both Claude and Copilot, so plugin-scope agents
  now carry only name/description/model/source_keys -- no tools: field,
  no Claude-only knobs (isolation, maxTurns, effort, memory,
  permissionMode).
- Root apm.yml registers all 7 marketplace packages (6 local plus
  mattpocock-skills as a remote entry) under versioning: per_package,
  matching this repo's existing independent-plugin-versioning practice.
- .claude-plugin/marketplace.json and every plugin's plugin.json are now
  apm-pack-compiled output, verified against the prior hand-maintained
  content: same names/descriptions/versions/licenses/authors, only
  cosmetic serialization differences (JSON key order, owner email vs.
  url, Unicode escaping).
- plugin-author and marketplace-author are retired now that apm-based
  authoring fully replaces their job; kyberforge bumped 1.3.1 -> 1.4.0
  for that removal, and the root marketplace catalog bumped
  0.3.1 -> 0.3.2 to match, per the version-bump convention now
  documented in apm-workflow's reference docs instead of a dedicated
  script (apm has no native version-bump automation).
- Fixed hardcoded pre-.apm/ path assumptions across
  .pre-commit-config.yaml, .pre-commit-hooks.yaml,
  scripts/check-scope-walkup-sync.sh, scripts/sync-vale-styles.sh,
  scripts/check-vale-style-sync.sh, six plugins' root plugin.json
  (stale skills/hooks/agents pointer fields that check-manifests.sh
  validates), and several tests/*.bats and tests/*.sh fixtures --
  including a bats REPO_ROOT relative-path depth bug (10 files, one
  extra .apm/ directory level to walk up) and a vale probe-path
  isolation regression introduced mid-fix.
- Corrected empirically-wrong assumptions surfaced this session in
  apm-workflow/apm-install's own reference docs: `apm marketplace
  package add` does not accept local paths (only owner/repo remote
  shorthand -- local packages are registered by editing apm.yml's
  marketplace.packages[] directly); `apm compile` is a consumer-side
  AGENTS.md/CLAUDE.md generator, not the plugin.json producer, and
  hard-fails on skill/agent-only packages without --clean; `apm plugin
  init <name>` nests a stray subdirectory when run with a positional
  name arg from inside a same-named directory; no native Copilot
  marketplace output profile exists; .mcp.json is merged into the
  compiled plugin.json content-aware and target-scoped, with no
  dependencies.mcp entry needed for simple passthrough; pipx is the
  correct pip fallback on externally-managed Python environments.
- Renamed agent-author's copilot.agent.md template asset to
  copilot.agent.md.template so apm compile's recursive *.agent.md glob
  stops misparsing the placeholder template as a real agent primitive.

Impact:
plugin.json and marketplace.json are compiled artifacts from here on --
editing them by hand is no longer the workflow; edit apm.yml/.apm/ and
run apm pack. CONTEXT.md's Plugin/Plugin marketplace glossary entries
reflect this. ADR-0001 is marked superseded, ADR-0006 moot, and
ADR-0010 updated for the new .apm/agents/ path (project/user scope
unaffected, per ADR-0016). Full local verification: claude plugin
validate --strict on all 6 plugins, apm audit --ci, apm marketplace
check, check-manifests.sh, and the full test suite (165/165 bats,
13/13 shell scripts) all pass clean.

Fixes: #90
Refs: #88, #89
ADR: 0015
ADR: 0016

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Ub96PyaSRD9BHPktotj1pC

2026-08-12 18:21:24 +00:00

5.6 KiB

Raw Blame History

name, description, compatibility, metadata, allowed-tools

name

description

compatibility

metadata

allowed-tools

pc-run

Use when the user wants to run pre-commit hooks, install git hooks, update hook versions, or maintain the pre-commit cache. Triggers on: "run pre-commit", "run all hooks", "check everything passes", "install hooks", "wire hooks into git", "update hook versions", "autoupdate", "bump revs", "clean the cache", "rebuild environments", "gc", "why is my hook failing", "hooks aren't running". Do not use for creating or editing `.pre-commit-config.yaml` — use `pc-author` for that.

Requires pre-commit installed and available on PATH.

Gotchas

Hooks not running on git commit almost always means pre-commit install was never run in this clone. Git hooks are per-clone — they are not committed to the repo.
When a hook modifies files (e.g. trailing-whitespace, end-of-file-fixer), the commit is blocked intentionally — the staged version is stale. The fix is git add -u && git commit. Do NOT call pre-commit install -f here; that is for overwriting existing hooks, not re-staging.
pre-commit autoupdate modifies .pre-commit-config.yaml in-place. Re-read the file after calling it to show the user the updated rev values.
The SKIP env var requires exact hook id values, comma-separated, no spaces: SKIP=check-yaml,gitleaks git commit -m "msg". A space after the comma silently skips nothing.
Never use git commit --no-verify (or -n) to bypass a failing hook. Hooks are the automated QA gate; bypassing them breaks the pipeline. Diagnose and fix the failure instead — see the hook-specific guidance below and in references/failure-patterns.md.
A stages mismatch — hook stage not installed — means the hook was added to the config but pre-commit install was not re-run with the correct -t flags. Hooks in stages not listed under default_install_hook_types will never fire.

Route

Determine intent from the user's request, then execute the matching operation:

User intent	Operation
"run", "check", "verify", "test hooks"	`pre-commit run --all-files` (default)
"staged", "simulate commit"	`pre-commit run` (staged files only)
"CI", "changed files only", "diff range"	`pre-commit run --from-ref <base> --to-ref <head>` — prefer this over `--all-files` on large repos
"install", "set up hooks", "wire into git"	`pre-commit install` — see Install
"pre-create environments", "install-hooks", "warm cache"	`pre-commit install-hooks` — see Install
"remove hooks", "uninstall", "tear down pre-commit"	`pre-commit uninstall`
"autoupdate", "update versions", "bump revs"	`pre-commit autoupdate`
"gc", "garbage collect"	`pre-commit gc`
"clean", "wipe cache", "rebuild from scratch"	`pre-commit clean` — see Clean

If the intent is ambiguous, default to pre-commit run --all-files.

Run

Default: pre-commit run --all-files. Never silently run staged-only.

pre-commit run --all-files

When hooks fail, read the output and:

Identify which hook failed and the specific cause. Be concrete: "gitleaks blocked config.json (high-entropy string on line 12)", not just "gitleaks failed".
Suggest a concrete next step. Common patterns are in references/failure-patterns.md.
Do NOT auto-fix code files. Do NOT modify .pre-commit-config.yaml. Those are the user's or pc-author's responsibility.

If the user asks to run only staged files: pre-commit run (no --all-files). If the user names a specific hook: pre-commit run <hook-id>.

Install

Only run when the user explicitly asks to install or set up hooks.

Before running, check for existing hook files:

ls .git/hooks/

If any hook files exist (e.g. a hand-written pre-commit), pre-commit install does NOT refuse or error — it defaults to migration mode, which runs the existing hook and pre-commit's hooks both. Only -f replaces the existing hook file outright, and that replacement is not reversible via pre-commit uninstall — uninstall only removes pre-commit from .git/hooks/, it does not restore whatever hand-written hook -f overwrote. If files are present, tell the user: "Existing hook files found at .git/hooks/<names>. Plain pre-commit install will run both; pre-commit install -f will overwrite them permanently instead. Proceed with plain install, or overwrite?" Wait for confirmation before using -f.

pre-commit install

Re-run with -t flags when default_install_hook_types was changed or when hooks in non-default stages aren't firing:

pre-commit install -t pre-commit -t pre-push -t commit-msg

To pre-create all hook environments without running hooks (useful for CI warm-up or first-time setup):

pre-commit install-hooks

To remove pre-commit from .git/hooks/ entirely:

pre-commit uninstall

Autoupdate

pre-commit autoupdate

After it completes, read .pre-commit-config.yaml and report which rev values changed. If the user wants to pin to exact SHAs (for reproducibility): pre-commit autoupdate --freeze.

Clean and GC

gc — removes only unused cached environments. Safe to run at any time:

pre-commit gc

clean — wipes the entire cache at ~/.cache/pre-commit. All hook environments will be re-downloaded on next run. Require explicit confirmation before running:

"This will wipe the entire pre-commit cache. All hook environments will be re-downloaded on next run. Proceed?"

Wait for the user to say yes before executing:

pre-commit clean

5.6 KiB Raw Blame History