# sentinel

> sentinel — milema-gif-sentinel. Use this tool when you need to prevent unintended edits to your codebase by implementing behavioral guardrails that ensure your agent reads before editing. It solves problems related to premature or unauthorized code modifications, providing a safe and controlled environment for development. With git integration, Sentinel accepts code repositories as input and outputs protected codebases with enforced read-before-edit protocols.

Canonical page: https://skillsregistry.net/skills/milema-gif-sentinel  
JSON: https://api.skillsregistry.net/v1/skills/milema-gif-sentinel

## Description

Behavioral guardrails for Claude Code — stop your agent from editing before it reads

## Trust

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

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **License:** MIT
- **Updated:** 2026-09-20

## Source

- **Source listing:** [GitHub](https://github.com/milema-gif/sentinel)

## 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": "milema-gif-sentinel"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/milema-gif-sentinel` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/milema-gif-sentinel/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