{"name":"Autoresearch: Autonomous Iteration for LLM-Powered Code Improvement","description":"Autoresearch is a powerful GitHub repository that transforms large language models like Claude Code, OpenCode, and OpenAI Codex into relentless improvement engines. Inspired by Karpathy's autoresearch, it enables autonomous, goal-directed iteration, following a \"Modify, Verify, Keep/Discard, Repeat\" loop to achieve compounding gains across various domains. This project offers a structured approach to automate development, security, and operational tasks with AI.","github":"https://github.com/uditgoenka/autoresearch","url":"https://osrepos.com/repo/uditgoenka-autoresearch","source":"osrepos.com","sourceDescription":"This repository profile is provided by osrepos.com, an open source repository discovery platform.","repositoryProfile":"https://osrepos.com/repo/uditgoenka-autoresearch","generatedFor":"open source discovery and AI-assisted research","markdown":"https://osrepos.com/repo/uditgoenka-autoresearch.md","json":"https://osrepos.com/repo/uditgoenka-autoresearch.json","topics":["ai","autonomous-agent","autoresearch","claude","github","automation","development","productivity"],"keywords":["ai","autonomous-agent","autoresearch","claude","github","automation","development","productivity"],"stars":null,"summary":"Autoresearch is a powerful GitHub repository that transforms large language models like Claude Code, OpenCode, and OpenAI Codex into relentless improvement engines. Inspired by Karpathy's autoresearch, it enables autonomous, goal-directed iteration, following a \"Modify, Verify, Keep/Discard, Repeat\" loop to achieve compounding gains across various domains. This project offers a structured approach to automate development, security, and operational tasks with AI.","content":"## Introduction\n\nThe `uditgoenka/autoresearch` repository introduces a groundbreaking skill for AI agents, particularly those powered by Claude Code, OpenCode, and OpenAI Codex. This project takes inspiration from Andrej Karpathy's original autoresearch concept, which demonstrated how a simple script could autonomously enhance machine learning models. Autoresearch generalizes this principle, turning any LLM into a \"relentless improvement engine\" capable of autonomous, goal-directed iteration. It operates on a core loop: modify, verify, keep/discard, and repeat, allowing for continuous, compounding gains in various tasks, not just limited to machine learning.\n\n## Installation\n\nGetting started with Autoresearch is straightforward, with options tailored for different platforms:\n\n*   **For Claude Code (Recommended via npx):**\n    bash\n    npx skills add uditgoenka/autoresearch\n    \n    After installation, restart Claude Code to access all 14 commands.\n\n*   **For OpenCode (Recommended via Guided Installer):**\n    bash\n    git clone https://github.com/uditgoenka/autoresearch.git\n    cd autoresearch\n    ./scripts/install.sh --opencode --global\n    \n    OpenCode commands use underscore naming, e.g., `/autoresearch_debug`.\n\n*   **For OpenAI Codex (Recommended via Guided Installer):**\n    bash\n    git clone https://github.com/uditgoenka/autoresearch.git\n    cd autoresearch\n    ./scripts/install.sh --codex --global\n    \n    Codex commands are invoked using mention syntax, e.g., `$autoresearch debug`.\n\nDetailed instructions and alternative manual installation methods are available in the official repository's README.\n\n## Examples\n\nAutoresearch offers a comprehensive suite of commands designed to automate various development and operational tasks. Here are a few highlights:\n\n*   **`/autoresearch` (Core Loop / Orchestrator):** The central command. In its classic mode, you define a goal, metric, and verification step, and it iteratively modifies code to improve the metric. The new autonomous orchestrator mode allows you to simply state a plain-language goal, and it intelligently selects and chains subcommands until the goal is met.\n    \n    /autoresearch\n    Goal: Increase test coverage from 72% to 90%\n    Scope: src/**/*.test.ts, src/**/*.ts\n    Metric: coverage % (higher is better)\n    Verify: npm test -- --coverage | grep \"All files\"\n    Iterations: 25\n    \n    Or, using the orchestrator:\n    \n    /autoresearch fix the login bug\n    \n\n*   **`/autoresearch:plan`:** Helps define the crucial Goal, Scope, Metric, and Verify steps by converting a plain-language goal into a validated configuration.\n    \n    /autoresearch:plan\n    Goal: Make the API respond faster\n    \n\n*   **`/autoresearch:debug`:** An autonomous bug hunter that iteratively finds and classifies bugs using falsifiable hypotheses. It can be chained with `/autoresearch:fix`.\n    \n    /autoresearch:debug\n    Scope: src/api/**/*.ts\n    Symptom: API returns 500 on POST /users\n    Iterations: 15\n    \n\n*   **`/autoresearch:fix`:** Designed to crush errors one-by-one, automatically detecting and prioritizing broken tests, types, or linting issues until the error count hits zero.\n    \n    /autoresearch:fix\n    Iterations: 20\n    \n\n*   **`/autoresearch:security`:** Performs a read-only security audit using STRIDE threat modeling, OWASP Top 10 sweeps, and red-team adversarial analysis.\n    \n    /autoresearch:security\n    Iterations: 15\n    \n\n*   **`/autoresearch:ship`:** A universal shipping workflow that guides any release (code, deployment, content) through 8 phases, with mechanically verifiable checklists.\n    \n    /autoresearch:ship --auto\n    \n\nThe project emphasizes \"8 Critical Rules\" for effective autonomous iteration, including bounded iterations, read-before-write, one change per iteration, mechanical verification, automatic rollback, and using Git as memory.\n\n## Why Use It\n\nAutoresearch stands out by providing a robust framework for leveraging LLMs to achieve continuous, measurable improvements. Its key advantages include:\n\n*   **Autonomous Improvement**: It automates the tedious cycle of modifying, testing, and refining, allowing agents to work tirelessly towards a defined goal.\n*   **Versatility**: Applicable across virtually any domain, from code optimization and bug fixing to security audits, documentation generation, and even product research, as long as a measurable metric or verifiable predicate can be established.\n*   **Efficiency**: The v2.1.0 architecture rebuild significantly reduced token usage, making operations more cost-effective and faster.\n*   **Safety and Reliability**: Features like automatic rollback, bounded iterations, and comprehensive hooks (e.g., `privacy-block`, `dangerous-cmd-block`) ensure that autonomous changes are controlled and safe.\n*   **Structured Workflow**: It provides a clear, rule-based approach to problem-solving, ensuring that progress is systematic and traceable through detailed TSV logs and Git commits.\n\n## Links\n\n*   **GitHub Repository**: [https://github.com/uditgoenka/autoresearch](https://github.com/uditgoenka/autoresearch){:target=\"_blank\"}\n*   **Author's Website**: [https://udit.co](https://udit.co){:target=\"_blank\"}\n*   **Karpathy's Autoresearch (Inspiration)**: [https://github.com/karpathy/autoresearch](https://github.com/karpathy/autoresearch){:target=\"_blank\"}\n*   **Claude Code Documentation**: [https://docs.anthropic.com/en/docs/claude-code](https://docs.anthropic.com/en/docs/claude-code){:target=\"_blank\"}\n*   **OpenCode**: [https://opencode.ai](https://opencode.ai){:target=\"_blank\"}\n*   **OpenAI Codex**: [https://developers.openai.com/codex](https://developers.openai.com/codex){:target=\"_blank\"}","metrics":{"detailViews":1,"githubClicks":0},"dates":{"published":null,"modified":"2026-08-24T19:41:31.000Z"}}