Skill Autoresearch

A disciplined take on iterating skills and workflow docs with actual benchmarks instead of vibes. You freeze a charter and evaluator first, baseline the current artifact, make one change at a time, and keep only measured improvements with append-only logs. The eligibility gate is smart: it can conclude "no ratchet justified" or "support surfaces drifted but the core is fine" instead of pretending every run needs mutation work. Works best when you have 3-5 representative prompts and deterministic checks. Explicitly scopes out GPU-bound ML experiments and hosted eval platforms like LangSmith or Promptfoo. If you have been rewriting SKILL.md files by feel and want git-friendly proof that edits actually helped, this is the structure.

skills.sh installs: 160
GitHub stars: 22
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add akillness/oh-my-skills --skill skill-autoresearch --agent claude-code

Installs into .claude/skills of the current project.

Files

Skill Autoresearch

Use this skill when the job is improving a reusable repo artifact with a frozen local benchmark.

The contract is simple: first decide whether a ratchet is justified at all, then freeze the judge, baseline the current artifact, change one meaningful thing only when needed, rerun the same harness, keep only measured improvements, and log both wins and reverts.

This skill is intentionally repo-local. It owns markdown/git-friendly ratchets for SKILL.md, SOPs, prompts, templates, and workflow docs. If the real need is product-scale traces, hosted dashboards, or app/runtime evaluation, route out instead of pretending this skill replaces LangSmith, Braintrust, Weave, or Promptfoo.

Read these support files before editing:

SKILL.md on GitHub

Sponsors

Promote your product hereReach 125k+ visitors monthly Promote your product hereReach 125k+ visitors monthly