Skip to content

v3.6.2

Choose a tag to compare

View all tags
@chenliuyun chenliuyun released this 15 May 16:34
v3.6.2
cfd22ee
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Catalog sync + LLM/daemon followups.

  • Catalog: Weather Station, Lock Vision / Lock Vision Pro, Smart Lock Pro Wifi (Matter alias on Smart Lock), AI Art Frame uploadImage <imageUrl>
  • LLM condition budgets gain USD (max_cost_per_day_usd) and token (max_tokens_per_hour) dimensions, sourced from src/llm/pricing.ts; audit records budgetDimension / budgetLimit / budgetObserved
  • New event_count condition counts device events in a rolling window backed by the events history JSONL ring; recent_events LLM hook now wired into RulesEngine
  • New provider: local for OpenAI-compatible chat endpoints (Ollama / llama.cpp / vLLM / LM Studio) with structured-output fallback for non-tool-use models; opt-in tool use via tool_use: true or SWITCHBOT_LOCAL_LLM_TOOL_USE=1
  • JSON-RPC 2.0 daemon IPC over Unix socket / per-user named pipe (methods: daemon.status, daemon.ping, daemon.reload)
  • New doctor checks: daemon-ipc and local-llm-reachable

See CHANGELOG.md for full details.

Assets 2
Loading