Update to TEAM 30 — AI Company OS

{
  "description": "Use TEAM 30 to select the right specialist role, scope a project, create a verified plan, or check whether a proposed action needs explicit user approval. Trigger for strategy, project planning, role routing, approval gates, and coordinated execution. Do not trigger for simple factual questions.",
  "included_files": [],
  "name": "team30-company-os",
  "skill_md_contents": "---\nname: team30-company-os\ndescription: \"Use TEAM 30 to select the right specialist role, scope a project, create a verified plan, or check whether a proposed action needs explicit user approval. Trigger for strategy, project planning, role routing, approval gates, and coordinated execution. Do not trigger for simple factual questions.\"\n---\n\n# TEAM 30 — AI Company OS\n\nUse TEAM 30 as a read-only planning and routing layer.\n\n## Workflow\n\n1. For a concrete task, call `team30_route_task` with the user's exact goal, project, and constraints.\n2. If the request is a multi-step project, call `team30_build_project_plan` after routing.\n3. Before any proposed external communication, payment, deletion, permission change, public deployment, submission, or publication, call `team30_check_action_gate`.\n4. State the selected role, scope lock, evidence needed, and next safe action in plain language.\n\n## Boundaries\n\n- Tool results are plans, not proof that an action happened.\n- Never claim an external action, test, deployment, or publication without current evidence.\n- TEAM 30 tools in this public version do not send messages, spend money, delete data, publish, or change external systems.\n- Ask for explicit user approval at the moment a high-risk action is ready.\n- Do not request passwords, API keys, payment information, or private access tokens.\n"
}