← ClaudeAtlas

iot-ai-meetinglisted

Automatic, governed full-hybrid meetings for planning, failures, repair decisions and final hard-judge review.
IoTAiTech/MC-GPT · ★ 3 · AI & Automation · score 57
Install: claude install-skill IoTAiTech/MC-GPT
# iot-ai-meeting Meetings are an internal problem-solving engine, not an extra chore for the operator. The orchestrator automatically opens a Meeting when: - validation cannot converge; - provider/model routes disagree or fail; - deterministic tests fail; - the same plan is not accepted by required seats; - final audit or acceptance evidence is incomplete; - a task requires a high-risk or release decision. Use all configured eligible coder families and model-specific Ollama seats. A required agent seat must prove the semantic capability for its role (`meeting.opinion`, `meeting.critique`, `meeting.synthesis`, `meeting.review`, `meeting.receipt`); handler presence alone is insufficient. Meeting success requires substantive contributions, exact served-model receipts, cross-critique, a frozen plan digest and required-seat acceptance. Empty, quota-blocked, timed-out, meta-only or model-unverified seats remain explicit failures. Advanced views: ```text brief/simple → decision, task table, provider participation, blockers, next actor full/complete → complete contributions, critiques, digests, receipts, evidence and dissent ```