Added Nov 4, 2026
computer-repair-skill
A Skill that has an AI agent gather evidence before making changes, to diagnose and repair Windows, macOS and Linux computers.
- Ecosystem
- Agent Skills
- Skill category
- Productivity
- Creator / organization
- 88lin
What it is
computer-repair-skill is a cross-platform computer diagnosis and repair Skill by GitHub user 88lin that installs into agents such as Codex, Claude Code and OpenClaw. It has the agent work like a careful repair engineer: read system state, logs and configuration first, judge the cause, and start with read-only diagnosis and inventory. For actions such as deleting, installing, elevating privileges, rebooting, partitioning and changing services, it gives the impact, rollback and verification plan first. The repository has 64 task-specific playbooks, loaded on demand by problem, covering network, performance, storage, app cleanup, data recovery and more.
Highlights:
- Reads system state, logs and configuration before judging the cause
- Diagnosis and inventory start read-only, and changes come with an impact, rollback and verification plan
- 64 task-specific playbooks loaded on demand
- Covers Windows, macOS and Linux
How to use:
- First, install the Skill into Codex, Claude Code or OpenClaw following the README.
- Then describe the computer problem in plain words, such as a full C drive or a slow machine.
- Finally, read the diagnosis and repair plan the agent gives and confirm before letting it act.
Who it's for:
People who want an AI agent to help troubleshoot and fix computer problems in a way they can check.
Before you start:
- You need an agent that supports Skills, such as Codex, Claude Code or OpenClaw, and the Skill itself needs no companion desktop app.
- The agent runs system diagnosis and repair commands that can change system configuration, so check the permissions and each action before it runs.
- If you use these features, weigh the safety risks yourself. AI Inspiration World is not responsible for any consequences.
- License: AGPL-3.0.
Based on the project's public documentation. Not tested by AI Inspiration World.