self-check
Skill7 starsChanged 5 months agoMITView on GitHub
Run local production/PR self-checks before pushing or submitting a PR. Mirrors pre-push checks and the GitHub PR CI checks where practical. Use when asked to "自检", "生产门禁自检", "跑门禁", "pre-push check", "PR checks", "before push", or "before PR".
- Installs packages
11 other projects use a version of this.
Adopted by
Most recently taken up by
- shakhruz/mila-coretheir own version, first seen Nov 1, 2026
- wizz92/skillstheir own version, first seen Oct 20, 2026
- xcota/postheir own version, first seen Oct 18, 2026
- borgr/paper-sharpenertheir own version, first seen Oct 17, 2026
- superzhangmch/claude-code-webtheir own version, first seen Oct 15, 2026
- Greg-Klein/implementation-factorytheir own version, first seen Oct 11, 2026
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- xcota/pos★ 32same name
- borgr/paper-sharpener★ 19same name
- Ron-dali/vibe-coding-rules★ 4same name
- xiding-code/job-application-skills★ 4same name
- shakhruz/mila-core★ 3same name
- fagemx/prismstack★ 2same name
- Greg-Klein/implementation-factory★ 1same name
- Greg-Klein/implementation-harness★ 1same name
- superzhangmch/claude-code-web★ 1same name
- wizz92/skills★ 1same name
1 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- release-publishSkill
- release-tagSkill
- release-versionSkill
- submit-prSkill
- sync-agent-skillsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
17 rules. 0 are common, 11 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 8 edited versions we read keep these.
11
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Before submit-pr or any request to push changes
- Distinctive: line 13
After committing feature, fix, refactor, test, docs, config, or release-prep changes
- Distinctive: line 14
When the user asks to run "自检", "生产门禁", "PR 检测", or "pre-push"
- Distinctive: line 15
When checking whether the current branch is ready for GitHub CI
- Distinctive: line 62
Check branch and working tree:
- Distinctive: line 74
If the working tree is dirty, continue only if the user asked to check the current local state.
- Very specific: line 82
Run exactly the commands from .husky/pre-push:
- Distinctive: line 100
Do not fix unrelated lint warnings during self-check.
- Distinctive: line 108
Run the frontend CI commands that are not already fully covered by pre-push:
- Distinctive: line 126
Run the backend CI commands:
- Distinctive: line 134
Prefer separate commands so failures are easy to identify.
- Distinctive: line 136
If local Linux system dependencies are missing, report the blocker and the CI packages required.
- Very specific: line 146
Use the run or verify skill if available, otherwise use the project commands:
- Distinctive: line 154
Only do this when the user asked for runtime verification or the change is UI/interaction behavior where tests/builds are not enough.
- Very specific: line 225
pnpm install --frozen-lockfilewould fail in CI: if lockfile/package changes are present, run or recommend pnpm install --frozen-lockfile … - Very specific: line 226
pnpm app:checkfails: run pnpm app:sync, inspect metadata changes, and commit them if expected. - Very specific: line 228
cargo fmt --checkfails: run cd src-tauri && cargo fmt, then review and commit formatting changes.
“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Self Check
Runs the repository's local production gate and the PR CI checks that can be reproduced locally. Use this before pushing a branch or opening a PR.
When to Use
- Before
submit-pror any request to push changes - After committing feature, fix, refactor, test, docs, config, or release-prep changes
- When the user asks to run "自检", "生产门禁", "PR 检测", or "pre-push"
- When checking whether the current branch is ready for GitHub CI
What PR Runs
GitHub PR checks are defined in .github/workflows/ci.yml and .github/workflows/ai-review.yml.
| PR Check | Local equivalent | Notes |
Frontend Checks → pnpm app:check | pnpm app:check | Verifies app metadata sync |
Frontend Checks → pnpm test | pnpm test | Vitest frontend/script suite |
Frontend Checks → pnpm build | pnpm build | Runs app:sync, tsc, and Vite build |
Backend Checks → pnpm app:check | Already covered by pnpm app:check | CI runs it again in backend job |
Backend Checks → pnpm build | Already covered by pnpm build | Builds frontend assets for Tauri context |
Backend Checks → cargo fmt --check | cd src-tauri && cargo fmt --check | Requires Rust toolchain |
Backend Checks → cargo clippy --all-targets --all-features -- -D warnings | Same command locally | Requires system Tauri/WebKit dependencies on Linux |
Backend Checks → cargo test --all-features | Same command locally | Requires Rust toolchain |
ai-review | Do not run locally by default | Requires PR context, GitHub token, and AI review secrets |
Local pre-push gate is defined in .husky/pre-push:
pnpm test
pnpm lint
pnpm format:check
pnpm test:tokensPrerequisites
- Clean working tree is preferred; if dirty, report the changed files before running checks.
- Dependencies are installed with
pnpm install. - Rust toolchain is available for backend checks.
- On Linux, backend clippy/tests may require the same system packages CI installs:
libwebkit2gtk-4.1-devlibgtk-3-devlibayatana-appindicator3-devlibrsvg2-devpatchelf
Workflow
Execute these phases in order. Report each phase result before moving on if it fails.
Phase 1 — Inspect Current State
Check branch and working tree:
git status --short --branchInspect recent commits when relevant:
git log --oneline -5- If the working tree is dirty, continue only if the user asked to check the current local state. Otherwise, ask whether to commit/stash first.
Output: branch name, clean/dirty state, and whether checks are being run against committed-only or working-tree changes.
Phase 2 — Run Local Pre-Push Gate
Run exactly the commands from .husky/pre-push:
pnpm test
pnpm lint
pnpm format:check
pnpm test:tokensA convenient single command is:
pnpm test && pnpm lint && pnpm format:check && pnpm test:tokensNotes:
pnpm lintmay report warnings. Warnings are acceptable if it exits successfully and reports0 errors.- Do not fix unrelated lint warnings during self-check. Report them separately if they are relevant.
Output: pass/fail for tests, lint, format, and token consistency.
Phase 3 — Run PR Frontend CI Equivalents
Run the frontend CI commands that are not already fully covered by pre-push:
pnpm app:check
pnpm buildNotes:
pnpm buildalready runspnpm app:sync, thentsc, thenvite build.pnpm testwas already covered in Phase 2; do not rerun it unless a previous command changed files.
Output: app metadata check and frontend build result.
Phase 4 — Run PR Backend CI Equivalents
Run the backend CI commands:
cd src-tauri && cargo fmt --check
cd src-tauri && cargo clippy --all-targets --all-features -- -D warnings
cd src-tauri && cargo test --all-featuresPrefer separate commands so failures are easy to identify. Avoid changing directories permanently; run commands with an explicit working-directory pattern when possible.
If local Linux system dependencies are missing, report the blocker and the CI packages required. Do not install system packages unless the user explicitly asks.
Output: Rust format, clippy, and test results.
Phase 5 — Optional Runtime Verification
For UI changes, run the app and observe the changed surface before claiming feature-level readiness.
Use the run or verify skill if available, otherwise use the project commands:
pnpm dev
# or
pnpm tauri:devOnly do this when the user asked for runtime verification or the change is UI/interaction behavior where tests/builds are not enough.
Output: what was launched, what user-visible behavior was observed, and any platform not available locally.
Phase 6 — Report
Report in this format:
Self-check result: PASS | FAIL | BLOCKED
Branch: <branch>
Working tree: clean | dirty (<summary>)
Checks:
- pre-push: PASS | FAIL
- pnpm test: ...
- pnpm lint: ...
- pnpm format:check: ...
- pnpm test:tokens: ...
- frontend PR CI: PASS | FAIL
- pnpm app:check: ...
- pnpm build: ...
- backend PR CI: PASS | FAIL | BLOCKED
- cargo fmt --check: ...
- cargo clippy ...: ...
- cargo test --all-features: ...
- ai-review: not run locally (<reason>)
Notes:
- <warnings, environment gaps, runtime verification notes>Quick Reference
Full local PR self-check:
pnpm test && pnpm lint && pnpm format:check && pnpm test:tokens
pnpm app:check
pnpm build
(cd src-tauri && cargo fmt --check)
(cd src-tauri && cargo clippy --all-targets --all-features -- -D warnings)
(cd src-tauri && cargo test --all-features)Pre-push only:
pnpm test && pnpm lint && pnpm format:check && pnpm test:tokensFrontend PR CI only:
pnpm app:check && pnpm test && pnpm buildBackend PR CI only:
pnpm app:check && pnpm build
(cd src-tauri && cargo fmt --check)
(cd src-tauri && cargo clippy --all-targets --all-features -- -D warnings)
(cd src-tauri && cargo test --all-features)Error Handling
pnpm install --frozen-lockfilewould fail in CI: if lockfile/package changes are present, run or recommendpnpm install --frozen-lockfilebefore PR.pnpm app:checkfails: runpnpm app:sync, inspect metadata changes, and commit them if expected.pnpm buildfails in TypeScript: fix type errors before rerunning.cargo fmt --checkfails: runcd src-tauri && cargo fmt, then review and commit formatting changes.cargo clippyfails: fix the reported warning/error; CI treats all clippy warnings as errors.- Linux Tauri dependency missing: report the missing package and the CI install list; ask before installing system packages.
- AI review not runnable locally: this is expected unless PR number, GitHub token, and AI review secret environment are available.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

