Research: harness-master
Cached research evidence for harness-master (not authority).
Quick Answer
Section titled “Quick Answer”Problem: Audit harness configs, usage/cost signals, and plugin, extension, MCP, or skill improvements. Use when tuning Claude, ChatGPT, Codex, Copilot, Cursor, Gemini, Grok Build, OpenCode, or Cherry. NOT for agents, MCP servers, or general app telemetry.
Stack / assumptions: Dry-run analysis first. Web access recommended for latest-doc checks. Optional Python 3.12+ for the surface discovery script.
Comparable alternative: A general-purpose agent instruction without a scoped skill contract
Repo summary:
Audit AI harness configuration quality, usage posture, and ecosystem options, then apply approved config fixes in the same session.
> Grounded in repository skills/harness-master/SKILL.md; treat as evidence, not authority.