This runs a structured PR review through seven distinct lenses including security, logic flaws, silent failures, and type design, with an extra "vibe security" scan for AI-generated code. It only surfaces findings with 76%+ confidence, which keeps the noise down. The workflow integrates with Jira, GitHub, GitLab, and Azure DevOps when available, posts findings as individual threaded comments instead of one wall of text, and includes a feedback loop that updates the skill definitions when blockers slip through. You get a three-option verdict at the end. It's opinionated about TDD and expects you to have a collection of framework-specific rules already configured.
npx -y skills add hoangnguyen0403/agent-skills-standard --skill code-review --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
github/awesome-copilot
alirezarezvani/claude-skills
microsoft/win-dev-skills
github/awesome-copilot