agentleFS
Sign inStart free
openJiuwen-aiopenJiuwen-ai/agent-core/.claude/skills/verification-loop/SKILL.md

verification-loop

Skill451 starsChanged 20 days agoApache-2.0View on GitHub

Formalizes agent-core's make check/type-check/test/fix pipeline into a structured 6-phase verification skill.

  • Commits and pushes

We split this skill into the rules it gives an agent

14 rules. 0 are common, 11 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 74 edited versions we read keep these.

  • 11

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill whenever you have completed a feature, fixed a bug, or are about to submit a PR.

    line 10
  2. Distinctive:

    Before every commit (or git push)

    line 15
  3. Distinctive:

    After any significant change spanning multiple modules

    line 16
  4. Very specific:

    After dependency changes (pyproject.toml, requirements*.txt)

    line 17
  5. Distinctive:

    Before marking a PR as ready for review

    line 18
  6. Distinctive:

    Run each phase in order. Stop immediately on failure.

    line 22
  7. Distinctive:

    Verify all modules compile without syntax errors:

    line 26
  8. Very specific:

    If any test fails, fix the test or the implementation before proceeding.

    line 56
  9. Very specific:

    Verify coverage stays above 80% for all modules under openjiuwen/core/ and openjiuwen/harness/.

    line 63
  10. Distinctive:

    No unintended changes (stale debug code, accidental formatting changes)

    line 84
  11. Distinctive:

    All new public APIs have corresponding tests

    line 85
  12. Distinctive:

    Commit message accurately describes the change

    line 86
  13. Distinctive:

    After running the full loop, report the result:

    line 90
  14. Distinctive:

    When skipping phases, state which ones were skipped and why in the verification report.

    line 134

“Common” is measured: kept by at least 6 in 10 of the 74 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Verification Loop

Runs the full verification pipeline before every meaningful commit. Use this skill whenever you have completed a feature, fixed a bug, or are about to submit a PR.

When to Run

  • Before every commit (or git push)
  • After any significant change spanning multiple modules
  • After dependency changes (pyproject.toml, requirements*.txt)
  • Before marking a PR as ready for review

6-Phase Verification

Run each phase in order. Stop immediately on failure.

Phase 1 — Build Check

Verify all modules compile without syntax errors:

python -m py_compile openjiuwen/

Phase 2 — Type Check

make type-check

Address all error: messages. warning: messages are acceptable but should be reduced over time.

Phase 3 — Lint

make check

Fix Ruff violations. If a rule is a false positive, add a per-line suppression comment, not a blanket disable in the config.

Phase 4 — Tests + Coverage

make test

If any test fails, fix the test or the implementation before proceeding. For @pytest.mark.slow tests, run selectively:

pytest -m "not slow"

Verify coverage stays above 80% for all modules under openjiuwen/core/ and openjiuwen/harness/. If coverage dropped, add tests for the missing paths.

Phase 5 — Security Scan

bandit -r openjiuwen/ -ll

Address all HIGH and CRITICAL findings. SECURITY-level findings require a code change, not a suppression. Only suppress with a documented reason.

Phase 6 — Diff Review

git diff --stat
git diff openjiuwen/

Review every changed line. Check for:

  • No unintended changes (stale debug code, accidental formatting changes)
  • All new public APIs have corresponding tests
  • Commit message accurately describes the change

Structured Verification Report

After running the full loop, report the result:

Verification Report
==================
Build:     PASS
TypeCheck: PASS (2 warnings)
Lint:      PASS (0 errors)
Tests:     PASS (142 passed, 3 skipped)
Coverage:  PASS (84.3% overall)
Security:  PASS (0 HIGH/CRITICAL)
Diff:      3 files changed, +47 -12
Status:    READY

Continuous Mode (CI)

In CI pipelines, run the full loop without stopping on the first failure. Collect all failures and report them together:

# CI mode: run all phases, report all failures at the end
set +e
make type-check 2>&1 | tee /tmp/typecheck.log
make check      2>&1 | tee /tmp/lint.log
make test       2>&1 | tee /tmp/test.log
bandit -r openjiuwen/ -ll 2>&1 | tee /tmp/bandit.log
set -e

# Report consolidated results

Partial Verification

For small, isolated changes, you may skip phases 1-2 if they are clearly unaffected. Use judgment:

Change typeCan skip phase 1-2?
Doc-only changeYes (phases 3, 4, 5, 6)
Test-only changeYes (phases 3, 4, 6)
Source code changeNo (all phases)
Dependency changeNo (all phases, especially 5)

When skipping phases, state which ones were skipped and why in the verification report.

Download now (ZIP)

  • Other versions80 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 80

    1. affaan-m/ECC★ 246ksame name
    2. Galaxy-Dawn/claude-scholar★ 5.9ksame name
    3. xu-xiang/everything-claude-code-zh★ 1.9ksame name
    4. loulanyue/awesome-claude-notes★ 272same name
    5. Resgrid/Core★ 235same name
    6. cloudnative-co/claude-code-starter-kit★ 153same name
    7. cohen-liel/hivemind★ 108same name
    8. tpavanipradeep/everything-claude-code★ 104same name
    9. MRCalderon3D/everything-game-dev-code★ 91same name
    10. mturac/everything-openai-codex★ 89same name

    70 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.

  • Used alongsideProjects with it also use everything-claude-code, everything-claude-code-conventions and 10 more.

    Other files people use alongside this one.

    Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
    • InfopibeSkill

      everything-claude-code

      Also in 10 projects that use this

    • mryeastcodeSkill

      everything-claude-code-conventions

      Development conventions and patterns for everything-claude-code. JavaScript project with conventional commits.

      Also in 8 projects that use this

    • InfopibeSkill

      investor-materials

      Create and update pitch decks, one-pagers, investor memos, accelerator applications, financial models, and fundraising materials. Use when the user needs investor-facing documents, projections, use-of-funds tables, milestone plans, or materials that must stay internally consistent across multiple fundraising assets.

      Also in 25 projects that use this

    • niloykumarbarmanSkill

      agent-sort

      Build an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.

      Also in 8 projects that use this

    • tpavanipradeepSkill

      nextjs-turbopack

      Next.js 16+ and Turbopack — incremental bundling, FS caching, dev speed, and when to use Turbopack vs webpack.

      Also in 15 projects that use this

    • niloykumarbarmanSkill

      product-capability

      Translate PRD intent, roadmap asks, or product discussions into an implementation-ready capability plan that exposes constraints, invariants, interfaces, and unresolved decisions before multi-service work starts. Use when the user needs an ECC-native PRD-to-SRS lane instead of vague planning prose.

      Also in 9 projects that use this

    • +1 more from Infopibe
  • In this repository6 other files it gives its agents.

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free