autoresearch

autoresearch is a Model training & evaluation skill for AI agents, published by pjhoberman in the autoresearch repository on GitHub. Skillspace rates autoresearch C (77/100) and it passed the safety scan; the rating is based on static analysis only.

Set up and run Karpathy-style autoresearch experiments on any codebase with a measurable metric. Use this skill whenever the user wants to autonomously optimize code by running iterative experiments — tuning search ranking, scoring functions, prompt templates, weight parameters, algorithm configurations, or any logic where changes can be evaluated against a numerical metric. Also trigger when the user mentions 'autoresearch', 'overnight optimization', 'autonomous experiments', 'autoresearch loop', 'Karpathy loop', or wants to 'let Claude Code optimize this while I sleep'. This skill generates the full experiment harness: instructions.md, eval script, test data template, and launch prompt — scoped to their specific codebase.

Grade

C

Quality

77 / 100

Safety

Passed

GitHub stars

21

Unproven: rated by static analysis only, not yet compared head-to-head.

Not launch-quality. It is graded C with a clean safety verdict, but our quality filters exclude it: it is a near-duplicate variant of another skill. So it is not in our launch-quality set, the skills our MCP server searches and recommends. How launch-quality is decided

Our two independent classifiers differ on this skill's category, so treat the placement as provisional.

This rating describes the version of the repository we crawled; it may have changed since, so go by its current files.

Skillspace indexes and rates this skill; it does not host or endorse it. Run third-party skills at your own risk.