self-repair

Automate AI-assisted repair and refactor workflows for the Crocbot codebase.

2|Updated Jan 27, 2026
One-click install
npx skills add https://github.com/moshehbenavraham/crocbot --skill self-repair-moshehbenavraham
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: self-repair
Source: https://github.com/moshehbenavraham/crocbot/tree/main/skills/self-repair
Command: npx skills add https://github.com/moshehbenavraham/crocbot --skill self-repair-moshehbenavraham

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables Crocbot to safely repair and refactor its own codebase using Claude Code (claude) as a high-thinking-budget sub-agent.

Core Features & Use Cases

  • Autonomous repairs: Initiates and guides code fixes or refactors across the repository with safety checks.
  • Code exploration & documentation: Analyzes project structure and documents findings in the identity workspace.
  • Controlled redeploy: Builds and restarts gateway after validating changes to ensure stability.

Quick Start

Run a repair pass to fix the latest failing module and redeploy the gateway after a successful build.

Frequently Asked Questions about self-repair

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate AI-assisted codebase repairs and refactoring?

Automating AI-assisted codebase repairs involves diagnosing, patching, building, and redeploying changes across your repository using a safe, auditable session lifecycle with explicit project root and workspace directory contexts.

What is the safest way to apply code refactors and redeploy changes?

The safest way to apply code refactors requires enforcing an explicit build process before redeploying any changes, ensuring structural improvements are validated and the gateway restarts only after a successful build.

How does an autonomous self-repair workflow handle code exploration and fixes?

An autonomous self-repair workflow analyzes the project structure, documents findings in the identity workspace, and initiates guided code fixes or structural improvements with built-in safety checks across the repository.

Can I use Claude Code as a sub-agent for repository refactoring?

Yes, you can use Claude Code as a high-thinking-budget sub-agent to safely repair and refactor its own codebase, guiding autonomous repairs and controlled redeployments after validating changes.

Do I need explicit directory contexts to run a self-repair session?

Yes, running a self-repair session requires explicit projectRoot and workspaceDir contexts to ensure the AI-assisted patching, building, and redeploying workflows remain auditable and safely scoped to the correct repository.

What limitations exist when automating codebase repairs with an AI assistant?

Limitations include the strict requirement to execute an explicit build before redeploying changes, meaning structural improvements or patches cannot bypass the validation phase or the controlled gateway restart lifecycle.