codex-autoresearch
CommunityAutonomous loop optimizer for Codex-driven work.
System Documentation
What problem does it solve?
Automates autonomous goal-driven experiments by running a deterministic modify–verify–keep/discard loop with Codex across a repository. Targets overnight and long-running optimization tasks in software projects and research contexts, supporting multiple modes and runtime workflows. Relies on frontmatter-driven discovery, structured artifacts for resume and audit, and robust preflight governance with rollback options.
Core Features & Use Cases
- Orchestrates single-iteration autonomic changes across a repository, with optional parallel execution.
- Supports seven modes (loop, plan, debug, fix, security, ship, exec) and a two-phase interaction model including heavy preflight checks and rollback governance.
- Reads and writes structured run artifacts (research-results.tsv, autoresearch-state.json, autoresearch-launch.json, autoresearch-runtime.json) and supports session resume and cross-run learning.
Quick Start
Describe your goal in plain language, then say go to start the autonomous loop.
Dependency Matrix
Required Modules
None requiredComponents
💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: codex-autoresearch Download link: https://github.com/BKZhao/MIMIC-Papers-Repro-Agent/archive/main.zip#codex-autoresearch Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.