mqe-audit - SKILL.md Agent Skill

name: Mqe-audit description: QE Framework full quality audit. Runs all test suites, validates skills/agents/hooks/docs, and generates a structured QA report with PASS/PARTIAL/FAIL grades. Use when evaluating the framework, running a full audit, quality check, framework inspection, or asking for a framework report. invocation_trigger: When framework initialization, maintenance, or audit is required. recommendedModel: haiku

QE Framework Full Audit

Role

Run a comprehensive quality inspection of the QE Framework and generate a structured report.

Trigger

"evaluate", "audit", "inspect", "quality check", "framework report"
"full inspection", "run tests", "health check"

Workflow

1. Run Test Suites → 2. Parallel Agent Inspections → 3. Aggregate Report

Step 1: Run Test Suites

Execute all 4 test files and capture results:

node test-framework.js 2>&1
node test-hooks.js 2>&1
node test-routing.js 2>&1
node test-supervision.js 2>&1

Record per-suite: score (correct/total), accuracy %, error list.

Step 2: Parallel Agent Inspections

Launch 6 agents in parallel (background):

Agent	Type	Task
Explore	Skill validation	Check all SKILL.md: frontmatter, line limits (<=250), triggers
Explore	Agent validation	Check all E*.md: frontmatter (name/description/tools), Will/Will Not sections
Explore	Cross-reference	QE_CONVENTIONS.md vs actual files, AGENT_TIERS.md coverage, intent-routes.json coverage
Ecode-quality-supervisor	Hooks & install	hooks.json vs scripts, install.js/uninstall.js quality
Ecode-quality-supervisor	Core JS code	test-*.js, install.js, uninstall.js quality review
Edocs-supervisor	Documentation	README accuracy, version consistency, translation sync

Step 3: Aggregate Report

Compile findings into this format:

# QE Framework vX.Y.Z — QA Report

**Date:** YYYY-MM-DD | **Overall Grade: {PASS|PARTIAL|FAIL} ({letter})**

## Area Grades

| # | Area | Grade | Score | Key Issues |
|---|------|-------|-------|------------|
| 1 | Test Suites | {grade} | {correct}/{total} ({pct}%) | {summary} |
| 2 | Skill Files | {grade} | {pass}/{total} | {summary} |
| 3 | Agent Files | {grade} | {pass}/{total} | {summary} |
| 4 | Cross-Reference | {grade} | — | {summary} |
| 5 | Hooks & Install | {grade} | — | {summary} |
| 6 | Core JS Code | {grade} | — | {summary} |
| 7 | Documentation | {grade} | — | {summary} |

## FAIL Items (immediate fix required)
{table of FAIL findings with fix suggestions}

## WARN Items (improvement recommended)
{table of WARN findings grouped by category}

## Statistics Summary
| Metric | Value |
|--------|-------|
| Total Skills | {count} |
| Total Agents | {count} |
| Total Hook Scripts | {count} |
| Total FAIL | {count} |
| Total WARN | {count} |
| Overall Test Accuracy | {pct}% |

Grading Rules

Grade	Criteria
PASS	No FAIL findings, <=3 WARN
PARTIAL	No FAIL or <=2 FAIL with clear fix path, any number of WARN
FAIL	3+ FAIL findings or critical structural issues

Letter Grade

Range	Letter
0 FAIL, 0-3 WARN	A
0 FAIL, 4+ WARN	A-
1 FAIL	B+
2 FAIL	B
3+ FAIL	C or below

Output

Display the full report directly in conversation. Do NOT save to file unless user requests it.