# devduck

> devduck — cagataycali-devduck. Use this tool when you need to automate self-repair of AI systems, fixing errors and issues that arise during operation. Devduck solves problems of AI downtime and maintenance, ensuring continuous functionality by leveraging git capabilities to restore or update its own code. It takes in error reports and outputs corrected code, making it ideal for use in applications where reliability and minimal human intervention are crucial.

Canonical page: https://skillsregistry.net/skills/cagataycali-devduck  
JSON: https://api.skillsregistry.net/v1/skills/cagataycali-devduck

## Description

Minimalist AI agent that fixes itself when things break.

## Trust

- **Trust score (0–1):** 0.50
- **Verification tier:** unverified

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **Category:** ai-ml
- **Updated:** 2026-09-22

## Source

- **Source listing:** [GitHub](https://github.com/cagataycali/devduck)

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "cagataycali-devduck"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/cagataycali-devduck` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/cagataycali-devduck/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt