smithery.ai

council-gate

Quality gate using LLM Council multi-model consensus for CI/CD pipelines. Use for automated approval workflows and pipeline quality checks.

First seen Apr 15, 2026

Installation

$ npx skills add https://smithery.ai

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery.ai · top by installs.

npx skills add https://smithery.ai

Browse all from smithery.ai

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Skill metadata

Parsed from SKILL.md frontmatter.

LicenseApache-2.0
Compatibilityllm-council-core >= 0.33, mcp >= 1.0, github-actions >= 2.0
Allowed toolsRead Grep mcp:llm-council/verify mcp:llm-council/audit
More metadata
category
ci-cd
domain
devops
council-version
0.33
author
amiable-dev
repository
https://github.com/amiable-dev/llm-council

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 4,200 B
  • docs SUMMARY.md 244 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 1 installs

SKILL.md

Council Gate Skill

Automated quality gate using multi-model consensus for CI/CD pipelines.

When to Use

  • Add AI-powered quality checks to GitHub Actions
  • Automate PR approval workflows
  • Gate deployments on multi-model verification
  • Enforce quality standards in pipelines

Exit Codes

Code Verdict CI/CD Behavior
0 PASS Pipeline continues
1 FAIL Pipeline fails
2 UNCLEAR Pipeline pauses for human review

Transcript Location

All deliberations are saved for audit:

.council/logs/{timestamp}-{hash}/
├── request.json      # Input snapshot
├── stage1.json       # Model responses
├── stage2.json       # Peer reviews
├── stage3.json       # Synthesis
└── result.json       # Final verdict

GitHub Actions Integration

name: Council Quality Gate

on:
  pull_request:
    branches: [main, master]

jobs:
  council-gate:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Install LLM Council
        run: pip install 'llm-council-core[http]'

      - name: Run Council Gate
        env:
          OPENROUTER_API_KEY: ${{ secrets.OPENROUTER_API_KEY }}
        run: |
          llm-council gate \
            --snapshot ${{ github.sha }} \
            --rubric-focus Security \
            --confidence-threshold 0.8

      - name: Upload Transcript
        if: always()
        uses: actions/upload-artifact@v4
        with:
          name: council-transcript
          path: .council/logs/

Configuration

Parameter Default Description
confidence_threshold 0.7 Minimum confidence for PASS
rubric_focus null Focus area (Security, Performance)
timeout 300s Maximum execution time
tier balanced Council tier (quick, balanced, high)

Output Schema

{
  "verdict": "pass",
  "confidence": 0.85,
  "blocking_issues": [],
  "rationale": "All models agreed...",
  "exit_code": 0,
  "transcript_path": ".council/logs/2025-12-31T...",
  "partial": false,
  "timeout_fired": false,
  "completed_stages": ["stage1", "stage2", "stage3"]
}

Under LLMCOUNCILSTRUCTUREDFINDINGS=true (ADR-051) the response also carries a typed findings array (all severities) and the verdict is computed as fail iff any critical finding exists. blockingissues is then the critical subset. Gate on the verdict / exitcode, not on blockingissues == [] (empty even on real FAILs on the default prose-scrape path).

Timeout Handling (ADR-040)

If timeoutfired: true, the gate timed out before completing all stages. This returns exit code 2 (UNCLEAR), pausing the pipeline for human review. Check completedstages to see how far it got. Consider using --tier quick for faster gate checks (the tier sets the timeout budget; there is no separate --timeout flag).

Example Usage

# Basic gate check
llm-council gate --snapshot $(git rev-parse HEAD)

# Security-focused gate
llm-council gate --snapshot $(git rev-parse HEAD) --rubric-focus Security --confidence-threshold 0.9

# Quick tier for faster feedback
llm-council gate --snapshot $(git rev-parse HEAD) --tier quick

Progressive Disclosure

  • Level 1: This metadata (~200 tokens)
  • Level 2: Full instructions above (~800 tokens)
  • Level 3: See references/ci-cd-rubric.md for CI/CD-specific scoring

Related Skills

  • council-verify: General verification
  • council-review: Code review with feedback