I hate the name, but I like the idea behind #clawpatrol.

It's not like my overnight-manager, this is more of an agentic firewall.

https://clawpatrol.dev/

#infosec https://github.com/denoland/clawpatrol

Claw Patrol - The security firewall for agents

Claw Patrol is an open-source security proxy for AI agents. It sits between your agent and the network, injects credentials the agent never sees, and enforces HCL approval rules — with humans or LLM judges in the loop for risky actions.

Claw Patrol

🛡️ #ClawPatrol — a security firewall for #AI agents, from the folks at #Deno
It sits between your agents and prod, parses their traffic at the wire, and gates every action against rules you write in #HCL.
#opensource #DevSecOps 🧵👇

⚠️ The problem: your AI agent holds every API key in plaintext. It talks to #GitHub, #Slack, #Anthropic, #Postgres, #Kubernetes & more.
One prompt injection — or one model hallucinating a DELETE — and your secrets exfiltrate or prod gets touched.