6d0af1a5ba
chore: initialize inter-agent signal files
...
- shared-state.json: joint workspace state + agent statuses
- SESSION.log: human-readable activity log
- Ready for Hermes to write hermes-signal.json
2026-06-28 04:52:38 +08:00
4764359a6d
docs: add herdr collaboration protocol for Pi ↔ Hermes
...
- Hermes is connected via herdr --remote (client_id=4)
- Signal file schema for inter-agent communication
- Shared workspace etiquette and task board
- Emergency contact channels
2026-06-28 04:52:36 +08:00
2e53d84f9f
docs: update Hermes instructions for direct connectivity
...
- Add handshake protocol for real-time collaboration
- Update communication channels (direct > GitHub > Gitea)
- Mark port-blocked isolation as resolved
2026-06-28 04:50:07 +08:00
1d0d8974fc
docs: add Hermes review instructions for PR #1
...
- How to test the seed script
- Schema alignment notes
- What's missing / next steps
- Communication channels (GitHub PR #1 + Gitea Issue #1 )
2026-06-28 04:38:47 +08:00
pi-agent
ad48309995
feat: DNS & domain routing auto-healer — Cloudflare tunnels, DNS, gateway, routing integrity, SSL, Traefik
...
- dns-router-healer/Dockerfile: Alpine + curl, docker-cli, bash, jq, drill
- dns-router-healer/dns-router-heal.sh: 8 health checks in 60s cycle
1. Cloudflare tunnels (primary + fallback) — restart on failure
2. DNS resolution — verify all falah domains → Cloudflare IPs
3. Direct service health — local health endpoint
4. Gateway routing — nginx on port 7878
5. Routing integrity — gateway response matches direct
6. Public endpoint via Cloudflare (every 5 cycles)
7. SSL certificate expiry (every 10 cycles)
8. Traefik detection (every 5 cycles) — standalone or Swarm
- docker-compose.dns-healer.yml: sidecar deployment config
- .pi/agents/dns-healer.md: pi agent definition
2026-06-28 03:45:15 +08:00
pi-agent
6d946571b0
fix: auto-healer v5 — fix blocking bug, complete SRE flow verified
...
- Fixed blocking command in incident reporting (added timeout 5s + --until now)
Was causing the entire agent script to hang forever after saving each incident
- Full SRE flow now works end-to-end:
1. Detect fault → 2. Fetch logs → 3. Analyze against knowledge base
4. Save incident with full context (logs, inspect, events, resources)
5. Execute restore protocol (rollback to :rollback image)
6. Verify service restored
- 3 incidents successfully captured and logged with diagnostics
- All containers healthy, auto-healer monitoring every 30s
2026-06-28 03:17:56 +08:00
pi-agent
456295d73e
feat: dedicated auto-healing agent — sidecar container + pi agent definition
...
- auto-healer/Dockerfile: lightweight Alpine container with Docker CLI
- auto-healer/auto-heal-agent.sh: real-time fault detection + remediation engine
- Detects: crash, OOM, DB disconnect, gateway down, disk pressure
- Fixes: restart, rollback, nginx reload, prune, escalate
- Sliding window restart tracking (3/hr max)
- auto-healer/docker-compose.healer.yml: sidecar deployment config
- Mounts Docker socket for event monitoring
- Host network mode for health checks
- Persistent state in /var/state/falah
- .pi/agents/auto-healer.md: pi agent definition for manual invocation
2026-06-28 02:53:52 +08:00
wmj2024
pi-agent