Building practical AI agent workflows with clear instructions, scoped execution, and verifiable results.
One vendor-neutral operating skill designed to help AI agents select relevant available capabilities and check their work.
- Cross-agent packaging: layouts for Codex, Claude Code, Gemini CLI, Antigravity, Cursor, Copilot, OpenCode, and generic Agent Skills hosts.
- Controlled installation: plan, install, verify, and rollback with ownership checks and retained snapshots.
- Verified builds: automated checks across Windows, Linux, and macOS with Python 3.12 and 3.13.
- Transparent scope: prerelease packages with documented limitations; native host activation is not yet certified.
Download releases · Read the guide · O'zbekcha qo'llanma
- Reusable skills and instructions for AI agents
- Practical automation with explicit permission boundaries
- Testing, reproducible packages, and clear documentation
Questions, bug reports, and suggestions are welcome through project issues. For sensitive security reports, use the project's private reporting channel.
Amaliy AI ish jarayonlari: aniq ko'rsatmalar, nazoratli bajarish va tekshirilgan natijalar.
