oskar/homelab-codex-ws
1
0
Fork 0
Code Issues Pull requests Actions Packages Projects Releases Wiki Activity
153 commits 19 branches 4 tags 1.1 MiB
Commit graph

6 commits

Author SHA1 Message Date
Oskar Kapala d2fb2b3d41 docs: onboard README + CLAUDE.md worktree discipline reminder 
scripts/onboard/README.md (new):
- Tool purpose and --node/--step/--from/--dry-run usage
- Full node.yaml field schema with annotations (ssh_user uid-1000
  gotcha, first_contact IP vs .local, deploy_autonomy/git_control gates)
- Step status table (00-access DONE, 00-preflight SCAFFOLD, 10-50 TODO)
- lib/ architecture: run() dry-run convention, yaml_get fallback caveats
- Gotchas/Learnings table from session

CLAUDE.md:
- Node Onboarding section: onboard.sh commands, pointer to README
- Multi-agent worktree mode: add explicit DISCIPLINE RULE — feature
  work must happen in agent.sh worktrees, not the main checkout;
  references the 2026-06-08 session that violated this

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-06-08 22:31:12 +02:00
Oskar Kapala 3b620ef7e3 docs(claude): multi-agent worktree mode section 
Main checkout = deploy-only. .agent-task marker triggers mandatory
loading of worktree-aware skill. Only the human runs scripts/dev/agent.sh.
 2026-06-03 17:41:35 +02:00
Oskar Kapala f381023206 docs(claude): add Definition of Done for services (smoke test + pytest) 
Lesson from brain-watchdog: code that was never run had a packaging bug
that caused a crash loop in production. New rule: docker build + short
smoke-run + pytest before any commit or deploy.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-06-01 20:38:39 +02:00
Oskar Kapala bf1415e4c1 feat(control-plane): route ha-diag-agent events through supervisor 
- 8 HA event types mapped to existing action types
- ha_websocket_dead → container_restart (homeassistant), 30-min cooldown
- 6 events → alert_only (entity_unavailable, integration_failed,
  automation_failing, update_available, recorder_lag,
  system_health_degraded), 1-hour cooldown
- ha_websocket_recovered → cancels matching pending container_restart
- state-aware suppression: skip HA events when homeassistant has an
  active containers_not_running incident < 5 min ago (avoids alert
  storms during HA restarts/updates)
- location_tag preserved through action pipeline for per-house
  telegram alerts
- executor: alert_only acknowledged as no-op success
- 18 tests covering all 8 event types, suppression, cooldown,
  dedup, location_tag, recovery cancellation
- CLAUDE.md: supervisor event routing table added

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-29 15:59:23 +02:00
Oskar Kapala f41ec5d0c5 docs: compress CLAUDE.md + fix zigbee2mqtt coordinator docs 
- CLAUDE.md: collapsed 5-section deployment block to single annotated
  block, removed inline emit_event signatures (kept path + type list),
  flattened runtime path tree to bullets, condensed node table note to
  reference capabilities.yaml, added CHELSTY docker-compose v1
  constraint; 156 → 113 lines (~750 → ~480 tokens)
- fix: zigbee2mqtt/README.md updated to TCP coordinator (SLZB-06U at
  192.168.1.105:6638, ezsp); removed stale /dev/ttyACM0 USB reference
  and corrected owner node from piha to chelsty-infra

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-29 14:17:23 +02:00
oskar 35e57cc789 docs(CLAUDE.md): update node model and override path convention 
- split CHELSTY into CHELSTY-INFRA and CHELSTY-HA in node roles table
- correct docker-compose override path to hosts/<node>/runtime/<service>/

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-05-20 15:27:46 +02:00