skilld
All skills
av avatar

/autoresearch

@7fb0398
by Ivan Charapanauav/skills16 stars
2

Runs autonomous keep/discard experiments on a codebase to optimize a single metric for a fixed duration, in the style of karpathy/autoresearch. Use when the user says "autoresearch" (optionally with a focus, e.g. "autoresearch the optimizer"), asks to run experiments on a repo overnight, to hill-climb or optimize a metric autonomously, or points at a repo with a karpathy-style program.md.

Ask your Agent

Use this Skill: https://skilld.dev/gh/av/skills/autoresearch

This session only. Nothing lands on disk.

README.md

≈174 tokens on demand. Your agent reads this file only when SKILL.md points to it.

<img src="../assets/logos/autoresearch.svg" width="72" align="left" hspace="12" alt="">

autoresearch

Runs autonomous keep/discard experiments on a codebase to optimize a single metric for a fixed duration, in the style of karpathy/autoresearch. Use when the user says "autoresearch" (optionally with a focus, e.g. "autoresearch the optimizer"), asks to run experiments on a repo overnight, to hill-climb or optimize a metric autonomously, or points at a repo with a karpathy-style program.md.

npx skills add av/skills --skill autoresearch

Part of av/skills — a library of agent skills for Claude Code, Codex, OpenCode and other coding agents.

Source: SKILL.md on GitHub

Third-party checks

1 warning1d3 checks · Risk SAFE
  • Gen Agent Trust Hub1d

    The autoresearch skill autonomously performs code experiments by generating changes and running shell commands. The primary security risk is Indirect Prompt Injection: the skill is designed to automatically adopt execution commands and configurations from a program.md file within the target repository. If this file is malicious, it could cause the agent to execute dangerous code. The skill also lacks sanitization for the ledger and log files it processes throughout the research loop.

  • Socket1d

    No alerts

  • Snyk1d

    Risk: MEDIUM · 1 issue

Provenance

Signed by skilld at 7fb0398. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last week.

Activeupdated last week

README badge

README badge for av/skills/autoresearch