Claude-skill-registry council-verify

install

source · Clone the upstream repo

git clone https://github.com/majiayu000/claude-skill-registry

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/council-verify" ~/.claude/skills/majiayu000-claude-skill-registry-council-verify && rm -rf "$T"

manifest: skills/data/council-verify/SKILL.md

Council Verification Skill

Use LLM Council's multi-model deliberation to verify work with structured, machine-actionable verdicts.

When to Use

Verify code changes before committing
Validate implementation against requirements
Check documents for accuracy and completeness
Get multi-model consensus on quality

Workflow

Capture Snapshot: Capture current git diff or file state (snapshot pinning for reproducibility)
Invoke Verification: Call
```
mcp:llm-council/verify
```
with isolated context
Receive Verdict: Get structured JSON with verdict, confidence, and blocking issues
Audit Trail: Persist transcript via
```
mcp:llm-council/audit
```

Parameters

Parameter	Type	Default	Description
`rubric_focus`	string	null	Focus area: "Security", "Performance", "Accessibility"
`confidence_threshold`	float	0.7	Minimum confidence for PASS verdict
`snapshot_id`	string	required	Git commit SHA for reproducibility

Output Schema

{
  "verdict": "pass|fail|unclear",
  "confidence": 0.85,
  "rubric_scores": {
    "accuracy": 8.5,
    "completeness": 7.0,
    "clarity": 9.0,
    "conciseness": 8.0
  },
  "blocking_issues": [...],
  "rationale": "Chairman synthesis...",
  "transcript_location": ".council/logs/..."
}

Exit Codes (for CI/CD)

```
0
```
: PASS - Approved with confidence >= threshold
```
1
```
: FAIL - Rejected
```
2
```
: UNCLEAR - Confidence below threshold, requires human review

Example Usage

# Verify current changes
council-verify --snapshot $(git rev-parse HEAD) --rubric-focus Security

# Verify specific files
council-verify --target-paths "src/auth.py,src/api.py" --snapshot abc123

Progressive Disclosure

Level 1: This metadata (~200 tokens)
Level 2: Full instructions above (~600 tokens)
Level 3: See
```
references/rubrics.md
```
for detailed rubric definitions

Related Skills

```
council-review
```
: Code review with structured feedback
```
council-gate
```
: CI/CD quality gate