Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLIAbout this skill
Design the evaluation harness for an agent loop — the proof layer that makes a loop trustworthy instead of merely confident. Use when asked to evaluate a loop, build an eval/verification harness or suite, measure how well a loop or agent system works, decide what to test, catch false completions, or grade a long-running agentic run. Lays out the 7-layer eval suite, makes false-completion-rate and repair-productivity first-class metrics, enforces deterministic-first-then-rubric, calibrates the judge, and keeps the regression harness repo-native. Delegates the deterministic gate to the contract's own verify scripts (with /verify-slice as an optional upgrade) — it designs the criteria, it does not build a new verification engine.
Quick start
- Copy the install prompt. Use the green button above this README.
- Send it to your agent. Paste it into your coding agent’s chat and send.
- Review the setup. Your agent checks the skill’s instructions, requirements, and installation scope.
Review, license & listing details
This is an imported listing. We haven’t certified its safety, license, or compatibility. Skills may include executable scripts.
Only SKILL.md was scanned. Supporting files and scripts were not checked. No safety grade has been assigned.
0 potential findings · Scanned 11/4/2026 · Scanner 0.2.1
Content SHA-256: 9299dcd9985d570df9c5c6d177f2c75efa97015f0b74c7bfce9f17709f810d58
About this listing
- Category
- General
- Reported license
- MIT
- Listing language (estimated)
- English
- Review status
- Not individually reviewed
- Agent compatibility
- Atlas installation supports Claude Code, Codex, and Cursor. Skill-specific dependencies and features are checked during setup.