Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This collection provides a structured set of skills designed to improve how AI agents handle complex technical tasks. By implementing specific protocols for requirement clarification, backend architecture, and library health audits, it moves agents from generic responses toward disciplined engineering practices.
The repository helps developers standardize agent workflows, ensuring that system designs are validated through test-driven development and that underspecified requirements are questioned before implementation begins.
WHO IT'S FOR
AI agent / automation builders
installing universal skills across multiple runtimes
Platform / DevEx teams
managing and auditing AI skill libraries
Backend / infrastructure engineers
designing APIs and database architectures with AI
Technical leads / architects
standardizing AI prompt best practices and requirements
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Make sure backend-development is installed in your agent before trying this task.
Use the backend-development skill to draft a concise REST API contract for [resource or feature]. Given [required operations], [authentication requirements], and [key fields], propose the endpoints, request and response shapes, pagination where relevant, and a consistent validation-error format. Keep this to a design review; do not modify application files.
Database engine: PostgreSQL, MySQL, or NoSQL context
Make sure database-design is installed in your agent before trying this task.
Use the database-design skill to review [schema or migration SQL] and [one important query]. Identify the three highest-impact issues involving normalization, constraints, indexes, or migration safety, then suggest focused improvements. Do not apply migrations or modify files.
Make sure verification-agent is installed in your agent before trying this task.
Use the verification-agent skill to verify this single claim or quotation: [claim]. Check it against [primary source URL or authoritative database] and [independent source URL]. Report the verdict, exact retrieved text, source URLs, UTC retrieval time, matching rule, content fingerprint, and independent cross-check so another person can reproduce the check. Clearly distinguish retrieval, source attribution, and factual truth.
Claude
Codex
Cursor
Gemini CLI
Copilot
Amp
OpenCode
Goose
Kilo Code