agentleFS
joris887joris887/exosuit/.claude/skills/quality-check/SKILL.md

quality-check

Skill9 starsChanged 3 months agoMITView on GitHub

Unified quality gate dispatcher. Runs selected quality checks with profile-aware defaults.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 6

    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. Very specific:

    If NO flags are provided, read the project profile from the Profile: line in CLAUDE.md and apply defaults:

    line 34
  2. Distinctive:

    If the profile is not set or not recognized, default to standard behavior.

    line 42
  3. Very specific:

    Before dispatching agent skills, run the quality-gate-sequence micro-component from .claude/prompts/quality-gate-sequence.md — this runs …

    line 46
  4. Distinctive:

    HARD GATE: If tests fail, stop here. Do not dispatch agent skills on failing code.

    line 48
  5. Distinctive:

    Run all selected checks in parallel for efficiency.

    line 60
  6. Very specific:

    If FAIL: list the specific blocking findings with file:line references so the caller can address them.

    line 94
  7. Distinctive:

    If a dispatched agent fails or times out: report that check as WARN with "agent failed — manual review recommended"

    line 98
  8. Distinctive:

    If no checks are selected and profile cannot be determined: fall back to standard defaults (code + tests + security)

    line 99
  9. Distinctive:

    If CLAUDE.md Commands has no test command configured: skip the test gate in quality-gate-sequence, note in report

    line 100

“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.

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

quality-check

Running quality gates: $ARGUMENTS

1. Parse Flags

Extract check selection from $ARGUMENTS:

FlagDispatches
--code/code-quality
--tests/test-validator
--security/security-audit
--architecture/architecture-check
--performance/performance-check
--allAll 5 checks above

2. Apply Profile Defaults

If NO flags are provided, read the project profile from the **Profile:** line in CLAUDE.md and apply defaults:

ProfileDefault Checks
lean--code (lint + complexity only)
standard--code --tests --security
strict--all plus integration-tester agent

If the profile is not set or not recognized, default to standard behavior.

3. Run Quality-Gate-Sequence First

Before dispatching agent skills, run the quality-gate-sequence micro-component from .claude/prompts/quality-gate-sequence.md — this runs lint, typecheck, and tests from CLAUDE.md Commands. These are fast, deterministic checks that catch issues before the more expensive agent analysis.

HARD GATE: If tests fail, stop here. Do not dispatch agent skills on failing code.

4. Dispatch Selected Checks

Dispatch each selected check as a parallel Task agent. Each invokes the corresponding skill:

For each selected check:
  Launch Agent (subagent_type: general-purpose) with prompt:
    "Run /<skill-name> on the current codebase. Return a structured report."

Run all selected checks in parallel for efficiency. Wait for all to complete.

<IF condition="Profile is strict AND --all was selected (explicitly or via default)"> Also dispatch the `integration-tester` native agent (`.claude/agents/integration-tester.md`) to independently verify acceptance criteria by running commands. </IF>

5. Unified Report

Combine results from all dispatched checks into a single report:

## Quality Gate Report

| Check | Status | Key Findings |
|-------|--------|-------------|
| Code Quality | PASS/WARN/FAIL | [summary] |
| Test Validator | PASS/WARN/FAIL | [summary] |
| Security Audit | PASS/WARN/FAIL | [summary] |
| Architecture | PASS/WARN/FAIL/SKIP | [summary] |
| Performance | PASS/WARN/FAIL/SKIP | [summary] |
| Integration Test | PASS/WARN/FAIL/SKIP | [summary] |

**Overall: PASS / WARN / FAIL**

Verdict rules:

  • PASS — all checks pass with no blocking findings
  • WARN — advisory findings only (no blocking issues)
  • FAIL — one or more checks have blocking findings

6. Return Verdict

Return the unified report and overall verdict. Callers (sprint-end, story-cycle) use this verdict for gate decisions.

If FAIL: list the specific blocking findings with file:line references so the caller can address them.

Error Handling

  • If a dispatched agent fails or times out: report that check as WARN with "agent failed — manual review recommended"
  • If no checks are selected and profile cannot be determined: fall back to standard defaults (code + tests + security)
  • If CLAUDE.md Commands has no test command configured: skip the test gate in quality-gate-sequence, note in report

Download now (ZIP)

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.