
skillcreatorai/ai-agent-skills
3 skills2.2k installs3.2k starsGitHub
Install
npx skills add https://github.com/skillcreatorai/ai-agent-skillsSkills in this repo
1Code ReviewCode Review is an agent skill that applies repeatable review patterns to pull requests and local changes so a solo builder can catch security, performance, maintainability, and testing issues before merge. It organizes feedback into security checks such as injection and hardcoded credentials, performance checks including N+1 queries and React re-render waste, code-quality signals like duplication and weak error handling, and testing gaps for new behavior and edge cases. Output follows a consistent summary template with severity buckets and suggested fixes at file and line granularity, which makes it easy to paste into GitHub, turn into tickets, or feed a fix loop. It fits indie teams who ship without a dedicated reviewer but still want audit-grade consistency. Invoke it during ship phase review, late-build hardening, or periodic operate-phase audits when you need the same checklist every time rather than ad-hoc chat opinions.965installs2Backend Developmentbackend-development is a reference-style agent skill for solo builders shaping server-side systems: RESTful routing examples, predictable JSON success and error payloads, and pragmatic Postgres-oriented schema guidance including UUID public identifiers, soft deletes, and indexing. The description directs use when you are designing APIs, database schemas, or broader backend architecture, and the body extends into microservices patterns and test-driven development for teams-of-one who still want production-grade conventions. Install it when your agent otherwise improvises inconsistent endpoint naming or ad-hoc error shapes across modules. It does not replace framework-specific docs for Rails, Nest, or Go routers, but it gives a shared contract your agent can apply while scaffolding CRUD APIs, planning migrations, or reviewing service splits before Ship-phase testing and security work.664installs3Code RefactoringCode Refactoring is an agent skill that teaches solo builders how to improve structure and readability without changing what the program does. It centers on classic triggers: after tests pass, before adding features, when smells appear, or when review asks for clarity—and on hard stops when coverage is missing, deadlines are unsafe, or the code is about to be deleted. The SKILL.md walks through long methods, deep nesting, and guard-clause rewrites with TypeScript snippets so an agent can suggest focused edits instead of vague “clean it up” advice. It fits Claude Code, Cursor, and Codex workflows where you already have a test harness or plan to add one. Use it as a checklist during Ship review or mid-Build hygiene, not as a license to rewrite entire modules without a safety net.593installs