a

autoharness

Autoharness is a self-learning skill layer designed for Claude Code, extracting and automatically updating skills from real conversations.

🌍 OverseasFreeOpen sourceDev开源Python自学习
Platforms: Desktop
Region
Overseas
Pricing
Free
Open source
Yes
GitHub Stars
★ 1.0k
Views
3
Source
GitHub
Added
2026-08-17
Last verified
2026-08-17

Overview

Autoharness is a self-learning skill layer for Claude Code. It extracts skills from actual user sessions and continuously updates and optimizes them during use. If a skill is no longer used, it is automatically removed, keeping the skill layer clean. Autoharness does not require separate data collection or replay loops; instead, it learns directly from users' work sessions. It merges skills from identical scenarios to avoid redundancy and only maintains skills it generates, without affecting other code. Autoharness is suitable for developers, especially those aiming to improve their efficiency with Claude Code.

Key features

  • Learn skills from real conversations
  • Automatically merge skills in similar scenarios
  • Update and remove skills based on usage
  • Maintain only self-generated skills

Use cases

Improve Claude Code usage efficiencyAutomate skill management and updatesReduce manual effort in maintaining skills

Pros

  • Automatic learning and skill updates
  • Reduction of duplicate skills
  • No additional configuration required

Limitations / notes

  • Requires internet connection
  • Depends on Python environment

Who it's for

DevelopersClaude Code users

This overview was compiled by AI from public sources and may contain inaccuracies — please refer to the official site.

FAQ

Is it free?

Yes, Autoharness is open-source and free.

Which platforms are supported?

Supports Linux and macOS platforms.

Do I need scientific internet access?

Not required, but internet access is needed to get the latest version.

Is Chinese supported?

Documentation and interface are primarily in English, with no explicit support for Chinese.

Similar agents

Something wrong? Let us know on the About page and we'll fix it.