Stores durable facts in a categorized, plain-markdown vault on disk, alongside your agent's built-in memory.
Coding
Huo15 Autoresearch Loop
Try it基于 Karpathy 自主研究循环的 OpenClaw Skill,Modify → Verify → Keep/Discard → Repeat forever
What it does
基于 uditgoenka/autoresearch(Karpathy 自主研究循环)的 OpenClaw Skill 实现。
The skill document
huo15-autoresearch-loop
基于 uditgoenka/autoresearch(Karpathy 自主研究循环)的 OpenClaw Skill 实现。
触发词
- 自动迭代
- autoresearch
- 跑起来别停
- 自动循环
功能
实现 Karpathy 的 Modify → Verify → Keep/Discard → Repeat forever 自主研究循环。
核心流程
- 用户说「自动迭代 [目标] [验证命令]」
- 初始化状态(目标/验证命令/迭代次数/范围)
- 循环:
- 修改代码/文件
- 调用验证命令
- 成功 → git commit + 记录
- 失败 → git revert
- 判断是否继续(迭代次数/收敛检测)
- 输出摘要
使用方式
自动迭代 [目标描述] [验证命令]
示例:
自动迭代 优化性能瓶颈 make test
自动迭代 修复所有lint错误 ./scripts/lint.sh
配置说明
config.json 控制循环行为:
{
"max_iterations": 50,
"verify_command": "",
"scope_globs": ["**/*.py", "**/*.js"],
"convergence_threshold": 3,
"commit_each_success": true,
"revert_on_fail": true
}
非侵入设计
- 只通过
exec调用脚本 - 不修改 OpenClaw 内核
- 状态持久化在本地文件
状态文件
每次迭代的状态保存在 ~/.openclaw/tmp/autoresearch-loop-state.json:
{
"goal": "优化性能",
"verify_command": "make test",
"iteration": 5,
"successes": 3,
"failures": 2,
"last_success": "2026-04-22T00:30:00Z",
"history": [...]
}
停止条件
- 达到
max_iterations - 连续失败超过
convergence_threshold - 用户发送「停止迭代」
- 收敛检测(连续 N 次成功)
version
1.0.0
Related skills
Save, search, and manage personal notes and knowledge bases in Get笔记 on explicit request.
Generate and edit Draw.io, Mermaid, and Excalidraw diagrams from natural language using a structured JSON spec.
Join a video meeting as an AI bot with voice, avatar, and screenshare across four operating modes.
Find why your productivity system keeps failing, then apply the smallest fix — capacity math, bottleneck routing, durable local notes.
Read and write Excel workbooks, worksheets, ranges, tables, and charts in OneDrive through Microsoft Graph with managed OAuth.
More from zhaobod1
Browse all skillsGenerate enterprise Word and native PDF documents across 39 format presets, with 7 contract subtypes.
Bootstrap a fresh OpenClaw workspace with a guided 4-step setup that writes five identity and preference files.
Adds Claude Code-style agent experience to OpenClaw without modifying the host.
麻省理工学院48小时学习法技能(青岛火一五信息科技有限公司)。完整还原 Ihtesham Ali 原始三问框架 + 反馈循环 + 完整 48h 三阶段时间线,叠加网上最佳实践(synthesis / contradictions / gaps / Feynman teach-back / weakness ana...
通过火山方舟Ark API调用Seedance 2.0生成第一人称带货短视频,v2 新增剧本驱动的配音(edge-tts/火山TTS)、背景音乐自动混音、字幕烧录,内置 8 套人设模板(传统女、时尚主播、老中医、厨房主妇、美妆博主、健身教练、户外探店、数码博主)。触发词:生成视频、带货视频、产品视频、拍视频、剧本...
规范 + 时尚的思维导图生成。输入 Markdown 大纲 / JSON / OPML / XMind,输出 XMind 2021+ (.xmind)、OPML、FreeMind (.mm)、Markdown、PNG、PDF、SVG;内置 13 种风格(modern / classic / dark / xiao...