RepoGuard
Architecture guardrail generator and auditor that keeps AI-generated code clean
Tested on Linux
Use it
Linux x86_64
install repoguard globally via npm install -g repoguard-rules
npm install -g repoguard-rules
Where the README and reality differ
- README: npm package is repoguard-rules with bin repoguard-rules / npx repoguard-rules
What worked: global install provides bin `repoguard` (and `repoguard-cli`), not `repoguard-rules`; launch is `repoguard mcp`
Add it to your agent
claude mcp add repoguard -- repoguard mcpJSON for Claude Desktop, Cursor and other clients
{
"mcpServers": {
"repoguard": {
"command": "repoguard",
"args": [
"mcp"
]
}
}
}Add repoguard mcp as an MCP server in your AI client, then call its tools such as repoguard_audit.
What it does
- Audit the architectural health of a codebaseTested
How to use it
- Connect your MCP client to the server started with
repoguard mcp. - Call the repoguard_audit tool with {"directory": "<project directory>"}.
- Read the text report it returns.
- Connect your MCP client to the server started with
- Generate AI context files tailored to your projectNot tested
- Audit uncommitted git changesNot shown working
- Find leaked secrets and credentialsNot shown working
How far pickfit got
Linux x86_64
- Installs
- Connects
- 1 of 4 features work
- Works offline
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
An MCP client called one of its tools
- Replayed from scratch
- Works offline
- Checked on 1 Nov 2026
- Version 1.7.0
An MCP client handshake over stdio listed 4 tools, and a call to repoguard_audit returned a result.
Differences the test machine hit
- README: npx repoguard-rules mcp launches server
In the test: installed globally works as `repoguard mcp` without npx -y fetch