Autoresearch
Autonomous Goal-directed Iteration. Apply Karpathy's autoresearch principles to ANY task. Loops autonomously — modify, verify, keep/discard, repeat. Supports optional loop count via Claude Code's /loop command. Invoking /autoresearch <free-form goal> builds a real-data benchmark harness, captures a baseline, and iterates with a regression gate until the goal is hit.
Install
npx skills add Muminur/autoresearch-skill-Andrej-Karpathy --skill autoresearchgit clone --depth 1 https://github.com/Muminur/autoresearch-skill-Andrej-Karpathy.git /tmp/autoresearch-skill-Andrej-Karpathy
mkdir -p ~/.claude/skills
cp -r /tmp/autoresearch-skill-Andrej-Karpathy/. ~/.claude/skills/autoresearchPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Claude Autoresearch — Autonomous Goal-directed Iteration Inspired by [Karpathy's autoresearch](https://github.com/karpathy/autoresearch). Applies constraint-driven autonomous iteration to ANY work — not just ML research. **Core idea:** You are an autonomous agent. Modify → Verify → Keep/Discard → Repeat. ## Subcommands | Subcommand | Purpose | |------------|---------| | `/autoresearch <goal>` | **Default path** — parse free-form goal, build harness, capture baseline, loop until goal met | | `/autoresearch` | Run the autonomous loop (default) | | `/autoresearch:plan` | Interactive wizard to build Scope, Metric, Direction & Verify from a Goal | | `/autoresearch:security` | Autonomous security audit: STRIDE threat model + OWASP Top 10 + red-team (4 adversarial personas) | ### Default Path: /autoresearch <free-form goal> When the user invokes `/autoresearch <goal>` with any free-form string after the command, parse the goal into seven slots, print the parsed-slot dump back for user visibility, then run the harness protocol in `references/benchmark-harness.md` before entering the loop. **Goal-parsing rubric:** | Slot | Extraction rule | Fallback | |------|-----------------|----------| | metric | First measurable noun (`latency`, `reliability`, `coverage`, `flakiness`, `bundle size`, `p95`, `accuracy`, `error-rate`, `LOC`, `build time`) | Ask user (1 sentence) |
Details
- Skill name
autoresearch- Repository
- Muminur/autoresearch-skill-Andrej-Karpathy
- License
- MIT
- Repo updated
- 19 April 2026
FAQ
What does the Autoresearch skill do?
Autonomous Goal-directed Iteration. Apply Karpathy's autoresearch principles to ANY task. Loops autonomously — modify, verify, keep/discard, repeat. Supports optional loop count via Claude Code's /loop command. Invoking /autoresearch <free-form goal> builds a real-data benchmark harness, captures a baseline, and iterates with a regression gate until the goal is hit.
How do I install the Autoresearch skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add Muminur/autoresearch-skill-Andrej-Karpathy --skill autoresearch. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Autoresearch popular?
It lives in Muminur/autoresearch-skill-Andrej-Karpathy, which has 20 GitHub stars and 2 forks.
Related skills
Browser Automation
19archubbuck/workspace-architect
Local Python-based browser automation toolkit using Playwright. Provides command-line tools for navigating, interacting with, and testing web applications without using MCP protocols. Supports clicking, typing, hovering, screenshots, content extraction, and JavaScript execution.
Notion Lifeos
24jiahao-shao1/notion-lifeos-skill
Notion LifeOS — PARA life management with Make Time journaling. Activate when: user mentions Notion, notes, tasks, projects, todos, journals, or uses phrases like 'take a note', 'add a task', 'what do I need to do', 'search my notes', 'jot down', 'today was great', 'record a thought', or Chinese equivalents like '帮我记一下', '加个任务', '今天要做什么', '查笔记'. Also: PARA method, personal knowledge management, daily reviews. Do NOT activate for: Notion API docs, Notion pricing/plans, workspace admin, database schema modification, or general productivity advice without data intent.
Travel Search Ru
19MissiaL/travel-search-ru
Use while planning a trip when Russian-catalog search is needed: package tours, hotels, flights, trains, excursions, prices, and booking links. Trigger on Russian requests: «спланировать путешествие», «подобрать тур», «найти отель/авиабилеты/жд билеты/экскурсии», «маршрут с актуальными ценами». Not for general advice without live search.
Xhs Notes Skill
22azlll/xhs-notes-skill
帮助创作者按单功能或工作流策划、生成、拆解和复盘小红书/XHS/RedNote 图文笔记;用户也可用“小红书图文笔记skill”“图文skill”“笔记skill”“小红书skill”等中文别名触发。适用于用户要做选题、笔记风格判断、封面制作建议、热点/热梗/热词/活动查询、违禁词查询、标签、评论钩子、封面分析、标题分析、内容分析、完整写笔记、复盘笔记或拆解笔记时。默认先温和顾问式拷问补齐上下文,再按意图执行单功能或工作流。
Be Ambitious Harness
23alexpiar/be-ambitious-harness
Harness for ambitious AI task statements. Use when the user writes /be-ambitious, "make this task deeper", "improve the task statement", "raise the bar", "strengthen the prompt" (or in Russian «сделай задачу глубже», «улучши постановку», «поставь амбициознее», «усили промпт») — or before launching any substantial task worth raising the bar on. Routes the task by scenario (design, data, code, writing, decision, generic), runs the analysis, builds a prompt with a gate check, hands it to a worker, runs an independent verifier. Not needed for five-minute tasks.
Trellis Workflow Enhancer
26lei1024/trellis-workflow-enhancer
Audit and apply real, project-local Trellis integrations for Matt Pocock and Waza skills, with phase routing, hook injection, durable artifact handoffs, and smoke-test evidence. Use when a user wants to strengthen, compare, modernize, integrate, or debug Trellis with Matt/Waza workflows, especially when skills are installed but do not actually trigger.
Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.