# TrustMesh

- **Event:** [Built with Opus 4.6: a Claude Code hackathon](https://cerebralvalley.ai/e/claude-code-hackathon)
- **When:** Feb 10 at 12:00 PM – Feb 17 at 10:00 AM (EST)
- **Where:** Location TBA
- **Team:** [Johnny Hung](https://cerebralvalley.ai/u/masterfung)
- **GitHub:** https://www.loom.com/share/b1ef90009d874d949f962a8497bbcabc
- **Demo video:** https://www.loom.com/share/28c45cbef27e4520ad05f348cf4ad0df
- **Gallery:** https://cerebralvalley.ai/e/claude-code-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/claude-code-hackathon/hackathon/gallery/208

TrustMesh is the trust layer for personal AI agents. Every person gets an encrypted vault and an Opus 4.6-powered agent. Agents gossip across trust boundaries — sharing what's needed, blocking everything else.

Peter asks his agent to prepare for Grandma Rose's visit. It searches his vault, then queries his wife Molly's agent through their shared family network. Molly's agent shares Rose's medications and care routine — but blocks her private journal and salary. Trust enforced
cryptographically.

When Rose has a car accident, the hospital issues a UCAN token: health data only, 30 minutes, full audit trail. Eight records released instantly. Family notified in real-time. Finances untouched.

15,000+ lines. 500+ tests. CLI + MCP server. 16-pod federation. One command to start. AI agents are coming — the missing piece is trust.

---

Markdown version of https://cerebralvalley.ai/e/claude-code-hackathon/hackathon/gallery/208. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt