Runtime profile preserved: claude_anthropic is intentional because this
smoke test explicitly invokes claude -p under each provider environment; a
direct_* profile would misdescribe the command chain.
Pings each Claude Code provider via claude -p with a trivial prompt and 90s timeout.
Telegrams Elmar only when one or more providers fail or time out — silent on full success.
Covers:
- anthropic (default for scheduler + persistent)
- glm (api.z.ai)
- minimax (minimaxi.com)
- kimi (api.kimi.com/coding/, canonical /models id kimi-for-coding, currently displays as K2.7 Code; K2.7/legacy K2.6 aliases are accepted but not the default)
Catches outages early so we know which alternates are usable before swapping the scheduler off Anthropic during a real incident.
python3 /home/lucienne/workspace/scripts/provider_smoke_test.py
Format: minute hour day month weekday
| Started | Status | Duration | Log |
|---|---|---|---|
| 10m ago | completed | 17s | View log |
| 1d ago | completed | 14s | View log |
| 2d ago | completed | 14s | View log |
| 3d ago | completed | 15s | View log |
| 4d ago | completed | 18s | View log |
| 5d ago | completed | 18s | View log |
| 6d ago | completed | 13s | View log |
| 7d ago | completed | 12s | View log |
Mission Control is the board for your delegated work: requests come in, Luci coordinates the next step, and evidence stays visible for review.
Luci is your always-on assistant for routing, status updates, and follow-through. Operators can still open deeper evidence when needed.