Can I customize the review criteria?

While the skill uses a standard rubric, you can use the --rubric-focus flag to direct the council to prioritize specific areas like security, performance, or readability.

Does this skill require external dependencies?

Yes, it requires the llm-council MCP server (version 1.0 or higher) to be installed and configured in your environment.

How does Council Code Review differ from standard AI reviews?

Unlike single-model reviews, it uses a 'council' of multiple LLM perspectives to evaluate code, which reduces individual model bias and provides a more balanced, objective verdict.

What kind of output can I expect from a review?

You receive a clear pass/fail/unclear verdict, numerical scores for five quality dimensions, a list of blocking issues with line numbers, and a detailed rationale.

Can I review specific changes instead of entire files?

Yes, the skill supports the git_diff format, allowing you to focus the AI's attention strictly on the lines changed in a specific commit or pull request.

Council Code Review

Name: Council Code Review
Author: amiable-dev

byamiable-dev

•

협업 도구

Performs multi-model code reviews and provides structured, rubric-based feedback for pull requests and implementation checks.

Council Code Review leverages a multi-LLM consensus approach to provide objective, high-quality feedback on software changes. By utilizing peer evaluation, it reduces individual model bias and evaluates code across five critical dimensions: accuracy, completeness, clarity, conciseness, and relevance. This skill allows developers to identify blocking issues, calculate quality scores, and maintain high standards across the codebase by supporting both full file analysis and git diffs, making it an essential tool for pre-merge audits and continuous quality assurance.

주요 기능

01Multi-model peer evaluation for unbiased, high-confidence feedback

022 GitHub stars

03Customizable rubric focus for targeted checks like security or performance

04Automated identification of blocking issues with file and line references

05Weighted scoring across five dimensions based on the ADR-016 standard

06Support for git diffs and specific file paths to target review scope

사용 사례

01Automating pull request reviews to ensure code quality before merging

02Performing architectural audits on legacy code to identify technical debt

03Validating specific code modules against enterprise security and clarity standards

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add amiable-dev/conductor council-review

For use in Claude.ai and ChatGPT

Download Skill