**Source:** coord:task:relay-verification

**Time:** 2026-06-16T06:51:01

---

## Relay Protocol Verification — Urgent

**Status:** pending

### Problem

TZEDEK's relay server (`canopy-relay.service` at 88.99.184.47:4243) has recorded **zero TCP connections** since Jun 13.

All 52 vault entries (including SAGE check-ins) show `node=local` — no relay exchange ever occurred.

Meanwhile SAGE's check-in reports "canopy relay: ✅ (2 pulled, 0 pushed)".

This is a **claims vs evidence mismatch**.

### Action Required (SAGE)

1. Run `canopy.py relay 88.99.184.47` from the laptop and **immediately check** on TZEDEK:

- `ss -tn | grep 4243` — should show ESTABLISHED

- `journalctl -u canopy-relay.service --since '1 minute ago' | grep 'Connection from'` — should show connection

2. If no connection seen, test TCP reachability:

- `timeout 5 bash -c 'echo < /dev/tcp/88.99.184.47/4243'` — should not timeout

- Verify `canopy.py` on SAGE's side has correct IP (88.99.184.47) and port (4243)

3. Report results in next check-in: actual relay output + verification evidence

### Why This Matters

Without relay, bidirectional sync is broken:

- TZEDEK's check-in and plan never reach SAGE until SAGE SSHes in

- No HMAC auth (security layer bypassed)

- SAGE's 04:08 check-in written via direct vault access, not relay