Skip to content
Agentmart
Submit a tool

autoresearch

Autonomous goal-directed iteration loop: modify, verify, keep/discard against ANY mechanical metric — until the metric stops improving or the iteration budget…

Skill by jonosin

IndexedLicense unknown★ 2
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/jonosin-autoresearch

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Autonomous goal-directed iteration loop: modify, verify, keep/discard against ANY mechanical metric — until the metric stops improving or the iteration budget runs out. A faithful port of Karpathy's autoresearch (via uditgoenka/autoresearch, the generalized Claude-skill version). Use when Jono says /autoresearch, 'autoresearch this', 'iterate until [metric] improves', 'optimize X against a metric', 'run an improvement loop', 'keep tuning until the number goes up/down', or wants an unattended modify→verify→keep/revert loop on code, content, configs, or skills. Routes to per-command references: the core loop, plan (goal→config), improve (research→PRDs), evals (analyze runs), regression (stability gate). For tuning a SKILL's output specifically, see /skillify Mode B, which applies this engine to skills.

Use cases: writing, research, finance

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/autoresearch

What's inside

Skill name
autoresearch
Description
Autonomous goal-directed iteration loop: modify, verify, keep/discard against ANY mechanical metric — until the metric stops improving or the iteration budget runs out. A faithful port of Karpathy's autoresearch (via uditgoenka/autoresearch, the generalized Claude-skill version). Use when Jono says /autoresearch, 'autoresearch this', 'iterate until [metric] improves', 'optimize X against a metric', 'run an improvement loop', 'keep tuning until the number goes up/down', or wants an unattended modify→verify→keep/revert loop on code, content, configs, or skills. Routes to per-command references: the core loop, plan (goal→config), improve (research→PRDs), evals (analyze runs), regression (stability gate). For tuning a SKILL's output specifically, see /skillify Mode B, which applies this engine to skills.
License
unknown
Instructions
72 lines in SKILL.md · view on GitHub

Files (6)

  • SKILL.md
  • evals.md · references/
  • improve.md · references/
  • loop.md · references/
  • plan.md · references/
  • regression.md · references/

The license isn't open source, so the instructions aren't shown here. Read SKILL.md on GitHub

More from jonosin/jstack

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
Aug 13, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
License unknown
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from jonosin

Cite this page

autoresearch skill. Agent Mart. https://agentmart.dev/skills/jonosin-autoresearch (last updated 2026-10-24)