ClawStrike
Project Concept
(The below is markdown, but might be rendered as plain text. Apologies if it comes across wonky wherever you happen to be reading it!)
How I Learned to Stop Worrying and Love Agents
I’ll be hacking on an in-progress project that protects Claws while they work. I’m hoping that by the end of the hackathon I can demo off an SRE bot that was able to work safely even while accessing scary production data.
Anyone can join in, especially with crazy ideas or examples of bots-gone-wild to test the concept out.
Demo Idea
- Run OpenClaw inside a container environment (we’re all doing that for security and isolation purposes, or tell me if I’m wrong!)
- Stand up a demo app that logs errors
- Tell OpenClaw how to get the logs, then to summarize the problem into a GitHub issue
- OpenClaw is gonna slurp in confidential data in the logs and include it in the issue (this is my expectation at least)
- See if another project I’ve been hacking on for AI workloads, Archodex, can be added to the OpenClaw environment to prevent sending confidential data
Entry
Status: Not Started
Team Roster
You must be registered for the event to view the team message board.
Chase Douglas Team Lead RSVP Approved
CEO at Archodex
Dhruvin Kamani RSVP Approved
Software Engineer at Softsmart Consulting