agentskillexchange Run autonomous improve verify keep-or-revert loops across coding tasks with autoresearch
Turn Claude Code, OpenCode, or Codex into a metric-driven loop that makes one change at a time, verifies it mechanically, and keeps or reverts automatically.
Prerequisites
Git repository, one supported agent environment such as Claude Code, OpenCode, or OpenAI Codex, mechanical verification command or metric
Installation
Requirements and caveats from upstream:
- Karpathy's autoresearch demonstrated that a 630-line Python script could autonomously improve ML models overnight — 100 experiments per night — by following simple princ...
- The wizard walks you through 5 steps: capture goal → define scope → define metric → define direction → validate verify command (dry-run). Every gate is mechanical — scope must resolve to files, metric must output a nu...
Basic usage or getting-started notes:
How It Works · Commands · Quick Start · Guides · FAQ
| /autoresearch | Run the autonomous iteration loop (unlimited) |
| Iterations: N | Add to inline config to run exactly N iterations then stop |
Extracted from upstream docs: https://raw.githubusercontent.com/uditgoenka/autoresearch/HEAD/README.md