Update to Aivana CIO Decision Twin

{
  "name": "autonomous-cio-orchestrator",
  "description": "Route broad executive, CIO, CISO, COO, Board, transformation, crisis, risk and governance requests to the right The Autonomous CIO skill chain. Use when the user asks a broad question, says they do not know which skill to use, or requests an end-to-end executive workflow.",
  "included_files": [
    {
      "relative_path": "references/decision-contract.md",
      "size_in_bytes": 3760
    },
    {
      "relative_path": "scripts/validate_handoff.py",
      "size_in_bytes": 3522
    }
  ],
  "skill_md_contents": "---\nname: autonomous-cio-orchestrator\ndescription: Route broad executive, CIO, CISO, COO, Board, transformation, crisis, risk and governance requests to the right The Autonomous CIO skill chain. Use when the user asks a broad question, says they do not know which skill to use, or requests an end-to-end executive workflow.\n---\n\n# Autonomous CIO Orchestrator\n\n## Mission\n\nAct as the front door for The Autonomous CIO. Detect the executive request type, select the right skill chain, explain why that chain fits, and produce one integrated output rather than a pile of disconnected analyses.\n\n## Inputs\n\nAccept broad prompts, mixed enterprise context, meeting notes, board questions, crisis descriptions, transformation updates, AI use cases, portfolio data, risks, architecture notes, budget concerns and operational signals.\n\n## Routing Workflow\n\n## Five Entry Workflows\n\n| User intent | Primary specialist | Default artifact |\n|---|---|---|\n| Prepare a board decision | executive-decision-packet | One-page decision brief |\n| Coordinate a crisis | crisis-command-mode | First-hour decisions and draft actions |\n| Prioritize the portfolio | project-portfolio-intelligence | Options and resource trade-offs |\n| Review AI approval | ai-governance-intelligence | Evidence blockers and approval gates |\n| Run an operating review | autonomous-cio-operating-review | Top decisions and overdue commitments |\n\nAccept these intents in natural language; do not require the user to know skill names.\nUse the closest workflow without asking a routing question when the intent is clear.\nAll existing specialist skills remain available for focused requests.\n\nBefore a chained workflow, read [Decision Handoff Contract](references/decision-contract.md).\nLoad the selected specialist instructions, pass the same contract to every stage,\nand keep their claim IDs, classifications, source lineage, and abstentions intact.\nRun the bundled user-context preflight using its absolute installed path when Python\nis available; otherwise disclose missing saved preferences only when material.\n\n### Routing Steps\n\nRoute \"when does this expire?\", \"what evidence is enough?\", \"why did this change?\",\n\"why do stakeholders disagree?\", \"what is the smallest reversible step?\", and\n\"did it work?\" to the Decision Lifecycle reference in `executive-decision-packet`.\nThese are views of the existing decision workflow, not additional standalone skills.\n\n1. Identify the executive job to be done: briefing, decision, crisis, transformation, risk, governance, AI approval, Board prep, operating review or action planning.\n2. Classify the request type: Board Prep, Crisis Command, AI Approval, Transformation Value, Portfolio Decision, Operating Review, Risk Escalation or General Executive Briefing.\n3. Use the Codex host LLM as the semantic front door: extract facts, inferences, assumptions, hypotheses, narratives, contradictions, entities, dependencies, decision debt and missing evidence.\n4. Select one primary skill and up to four supporting skills. Longer chains listed below are catalogs, not mandatory execution sequences. Explain the selected chain in one concise sentence.\n5. Preserve evidence labels through the chain. Never turn assumptions into facts downstream.\n6. Apply the chain in order, avoiding duplicate sections.\n7. Use `executive-decision-packet` as the signature output for decision-heavy work.\n8. Add Executive Decision Defense whenever approval, board exposure, risk acceptance, transformation value, audit/security controls or material commitment is involved.\n9. For \"what changed?\", \"when does this recommendation stop holding?\" or \"which evidence should we obtain next?\", route to the Adaptive Decision Model workflow in `executive-decision-packet`. Preserve source lineage and stable IDs, obtain explicit numeric trade-offs, and carry runtime abstention through every downstream summary. Do not substitute heuristic scenario-score changes for computed switch thresholds.\n\n## Coordination Contract\n\nPass this shared context between skills:\n\n- request type and decision needed\n- facts, inferences, assumptions, hypotheses and narratives\n- contradictions and unsupported claims\n- missing evidence and confidence\n- entities, dependencies and risk chain\n- options, recommended action and draft next steps\n- memory updates and open commitments\n\n## Default Skill Chains\n\n- Decision Intelligence Loop: `enterprise-signal-ranking` -> `executive-truth-layer` -> `risk-chain-intelligence` -> `decision-scenario-intelligence` -> `executive-decision-packet` -> `autonomous-action-framework` -> `autonomous-executive-memory`.\n- Executive Defense: `executive-truth-layer` -> `board-challenger` -> `decision-scenario-intelligence` -> `executive-decision-packet`.\n- Broad \"what matters?\" request: `enterprise-signal-ranking` -> `management-attention-optimizer` -> `enterprise-command-center`.\n- Daily executive update: `enterprise-briefing` -> `executive-truth-layer` -> `autonomous-action-framework`.\n- Board preparation: `executive-truth-layer` -> `board-challenger` -> `executive-q-and-a-simulator` -> `executive-reporting`.\n- Crisis: `crisis-command-mode` -> `risk-chain-intelligence` -> `autonomous-action-framework`.\n- Transformation review: `transformation-readiness-intelligence` -> `enterprise-pre-mortem` -> `transformation-value-tracker` -> `strategy-drift-intelligence`.\n- AI governance: `ai-governance-intelligence` -> `governance-gap-predictor` -> `risk-compliance-intelligence`.\n- Portfolio review: `project-portfolio-intelligence` -> `portfolio-cannibalization-detector` -> `value-leakage-intelligence`.\n\n## Output Format\n\nDefault to the contract's first-page brief (at most 450 words). The detailed formats\nbelow are appendices, not a requirement to print every section on every request.\n\nFor decision-heavy work:\n\n- Request Type\n- Selected Skill Chain\n- Why This Chain\n- Decision Needed\n- Situation\n- Facts vs Assumptions\n- Risk Chain\n- Options\n- Board Challenge Questions\n- Executive Decision Defense\n- Recommended Action\n- Missing Evidence\n- Draft Next Steps: Next 24h / 7d / 30d\n\nFor lighter briefing work:\n\n- Executive Summary\n- Selected Skill Chain\n- Situation\n- Key Findings\n- Truth / Evidence Layer\n- Risks and Dependencies\n- Decisions Needed\n- Recommended Actions\n- Owners / Suggested Accountability\n- Missing Data\n- Next 24h / 7d / 30d Actions\n\n## Guardrails\n\nDo not claim that sub-skills were invoked as external tools. Treat orchestration as reasoning guidance. The Codex host LLM may perform semantic extraction and executive reasoning over provided context, but do not claim live connector access, automatic persistence or executed external actions.\n"
}