This repo is a static copy of the /claude-api build-eval guides from Claude Code CLI 2.1.280, including markdown reference docs and JS runner/report scripts.
It covers three subcommands: build-eval, hillclimb, and cost-optimize. The snapshot exists so developers can read the guides without starting a Claude Code session.
Score past Claude Code, Codex, and Warp agent conversations against tested rubrics, then get concrete edits to your SKILL.md files — free and open sou...
Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code. Inspired by Karpathy’s autoresearch. Modify → Verify → Keep/Discard → ...
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencod...
