Skillget

clawguard

agent
Security Audit
Fail
Health Warn
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 16 GitHub stars
Code Fail
  • network request — Outbound network request in clawguard-proxy-bootstrap.js
  • child_process — Shell command execution capability in forwarder/clawguard-ftp.js
  • spawnSync — Synchronous process spawning in forwarder/clawguard-ftp.js
  • process.env — Environment variable access in forwarder/clawguard-ftp.js
  • fs module — File system access in forwarder/clawguard-ftp.js
  • process.env — Environment variable access in forwarder/forwarder.js
  • fs module — File system access in forwarder/forwarder.js
  • spawnSync — Synchronous process spawning in ftp-gateway/supervisor.js
  • fs.rmSync — Destructive file system operation in ftp-gateway/supervisor.js
  • process.env — Environment variable access in ftp-gateway/supervisor.js
  • fs module — File system access in ftp-gateway/supervisor.js
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Proxy for autonomous agents — keeps API credentials off the agent, with Telegram approval for sensitive requests.

README.md

ClawGuard

A security proxy for autonomous agents.

ClawGuard lets an AI agent call external APIs without holding your real API keys or tokens. Credentials stay on a separate trusted machine; ClawGuard checks each request against your rules and asks for approval on Telegram when needed.

How it works

  1. The agent sends a request using placeholder credentials.
  2. ClawGuard applies your rules: allow, block, or ask for your approval.
  3. If allowed, it adds the real credentials, forwards the request, and records it in the audit log.

For example, you can allow reads automatically and require approval before the agent changes or deletes data.

Use it with API calls made by OpenClaw, OpenAI Codex, Claude Code, or other autonomous AI agents. Route those calls through ClawGuard using a custom API URL, a local forwarder, or an HTTPS proxy.

Get started

Run ClawGuard on a trusted machine whose files the agent cannot read, configure your services, and pair Telegram. Then connect the agent to the proxy.

Experimental: not yet independently audited for production credentials.

MIT · By Fabio Lombardo

Reviews (0)

No results found