A factory for building production-grade agent skills — benchmarked against gold standards, autonomously improved via an autoresearch loop, and verified by independent multi-agent consensus.
2 files for Codex, OpenCode and Claude Code: agent-skills-harness AGENTS.md, agent-skills-harness CLAUDE.md — 1,925 tokens loaded in every session.
AGENTS.md C 1,922 tok CLAUDE.md A 3 tok These files are a-tokyo/agent-skills-harness's own configuration — they tell Codex, OpenCode and Claude Code how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.