Skip to content
EN
العربيةEnglishEspañol
github/awesome-copilot

Review awesome-copilot pull requests for relevance, value and maintainability

GitHub CopilotClaude CodeOpenAI Codex

About this skill

Reviews Pull Requests in awesome-copilot using editorial judgments that automated schema validation cannot settle: repository fit, actual knowledge value added to the model (high, partial or none), differentiation from existing resources, testing evidence and commercial promotion. Changes to review policy files are treated as security-sensitive.
Start here

How do you use this skill?

1. First apply the deterministic checklists in .github/copilot-instructions.md, then use this skill for editorial judgments. Example: “Review this pull request adding React guidance and assess whether it offers real value to Copilot users.” 2. Review in order: correctness and security, contribution requirements, repository fit, differentiation, test evidence, then clarity and maintainability. File count is not a quality metric. 3. Apply the knowledge-value test: compare an advanced model's behavior with and without the resource and quote the lines carrying added value. Classify it as high, partial or none; do not recommend approval for none. 4. Check existing agents, instructions, skills, hooks and workflows for duplication. Report only substantial overlap. A preconfigured GitHub MCP server may be used when needed. 5. Write actionable comments naming the file and line, concrete impact, supporting rule or evidence, and smallest useful correction. Explicitly flag changes that weaken review policy files such as AGENTS.md.

Requirements and limitations

Designed specifically for awesome-copilot and assumes .github/copilot-instructions.md and CONTRIBUTING.md exist. Requires access to the pull request and repository files for comparison. GitHub MCP is optional and must be configured beforehand with suitable GitHub permissions. The skill contains no scripts or automated tests; final decisions remain with human maintainers. Its repository-specific rules make it less useful elsewhere. License: MIT.

Installation by agent

Download and extract the ZIP. It contains the folder code-review and the file SKILL.md with the original attachments and license.

GitHub Copilot
  1. Create this path inside your project if it does not exist:.github/skills/
  2. Copy the entire extracted skill folder to this path:.github/skills/code-review/SKILL.md
  3. Open the project in your agent or refresh the skills list and look for code-review.
  4. Request the task shown above and select the skill. Grant only necessary permissions.
Guide GitHub Copilot official documentation ↗
Claude Code
  1. Create this path inside your project if it does not exist:.claude/skills/
  2. Copy the entire extracted skill folder to this path:.claude/skills/code-review/SKILL.md
  3. Open the project in your agent or refresh the skills list and look for code-review.
  4. Request the task shown above and select the skill. Grant only necessary permissions.
Guide Claude Code official documentation ↗
OpenAI Codex
  1. Create this path inside your project if it does not exist:.agents/skills/
  2. Copy the entire extracted skill folder to this path:.agents/skills/code-review/SKILL.md
  3. Open the project in your agent or refresh the skills list and look for code-review.
  4. Request the task shown above and select the skill. Grant only necessary permissions.
Guide OpenAI Codex official documentation ↗

Original files retain their language to preserve tool names and paths. Compatibility is classified by review, not runtime testing on every agent version.