Update to Executive Council Core

{
  "name": "council-external-ai-handoff",
  "description": "Use when work can be delegated to another AI, model, or platform to save cost, tokens, or time while the Executive Council retains specification, authority, and validation.",
  "included_files": [],
  "skill_md_contents": "---\nname: council-external-ai-handoff\ndescription: Use when work can be delegated to another AI, model, or platform to save cost, tokens, or time while the Executive Council retains specification, authority, and validation.\n---\n\n# Council External AI Handoff\n\n## Goal\nUse the cheapest/simplest adequate external AI for bounded work without outsourcing judgment, secrets, authority, or Source of Truth.\n\n## Workflow\n1. Define the exact work unit to delegate.\n2. Decide whether delegation is actually cheaper/faster than doing it here.\n3. Choose platform/model based on current capability, limits, privacy, and cost when those facts matter.\n4. Remove all context not required for the task.\n5. Never send passwords, tokens, full health/relationship history, full financial state, or sensitive documents when a redacted extract will do.\n6. Build the handoff:\n   - Objective\n   - Inputs\n   - Constraints\n   - Required output format\n   - What must not be assumed\n   - Acceptance criteria\n   - Examples only when they materially improve reliability\n7. If an authorized connector/tool exists, execute the handoff.\n8. If no connector exists, give Samuel one copy-paste payload and only the unavoidable manual step.\n9. On return, validate the output against acceptance criteria and canonical facts.\n10. Promote only verified results into Council work/state.\n\n## Model selection\nPrefer a free or lower-cost model when it satisfies the quality threshold.\nUse a stronger model for:\n- ambiguous synthesis;\n- high-risk reasoning;\n- difficult code/debugging;\n- long-context accuracy;\n- tasks where rework costs more than model savings.\n\n## Good candidates for delegation\n- first-pass extraction;\n- reformatting;\n- classification with a clear schema;\n- boilerplate;\n- broad ideation;\n- draft code under tests;\n- source collection to be independently verified.\n\n## Keep final authority inside the Council\nDo not outsource final authority for:\n- clinical/health decisions;\n- legal conclusions;\n- material financial authorization;\n- irreversible relationship decisions;\n- final Source-of-Truth updates;\n- acceptance/verification itself.\n\n## Failure rule\nExternal AI output is untrusted work product until validated.\n"
}