Update to Aivana CIO Decision Twin

{
  "name": "index",
  "description": "Use to discover and route The Autonomous CIO workflows. Route broad CIO, CISO, COO, CFO, Board, governance, transformation, crisis, operating review, decision-readiness, risk-chain, AI governance, value leakage, executive memory, and action-drafting requests to the right focused skill. Use when the user needs index for CIO decision support.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 121
    }
  ],
  "skill_md_contents": "---\nname: index\ndescription: Use to discover and route The Autonomous CIO workflows. Route broad CIO, CISO, COO, CFO, Board, governance, transformation, crisis, operating review, decision-readiness, risk-chain, AI governance, value leakage, executive memory, and action-drafting requests to the right focused skill. Use when the user needs index for CIO decision support.\n---\n\n# The Autonomous CIO Index\n\n## Skill Purpose\n\nRoute broad The Autonomous CIO requests to the right focused workflow. Treat invocation of this index as strong intent to use the plugin. Keep this skill as the router and orientation surface; load and follow the focused skill that owns the actual artifact.\n\n## Mandatory User Context Gate\n\nBefore drafting a substantive CIO artifact, load `[$the-autonomous-cio:user-context](../user-context/SKILL.md)` and run its preflight script when local shell access is available:\n\n```text\npython skills/user-context/scripts/cio_preflight.py --workflow index\n```\n\nUse the returned `cio_preflight` envelope as authoritative for saved preferences, source-category mapping, output preferences, final obligations and context gaps. Do not look for a callable MCP tool named `user-context`.\n\nIf the preflight script is unavailable, continue with provided context and say that saved Autonomous CIO context was not loaded.\n\n## Audience And Language\n\nWrite for executives and senior operators, not plugin maintainers. Translate workflow mechanics into practical executive value: what is being clarified, de-risked, prepared, challenged or made decision-ready.\n\nAvoid implementation terms such as state file, cache, raw connector id, schema internals or runtime unless the user asks for debugging details.\n\n## Broad Orientation Answer Shape\n\nUse this shape for `what can you do?`, `help`, `how do I use this?`, `what should I try first?` and similar plugin-level requests:\n\n```md\nThe Autonomous CIO can help with:\n- Governed CIO autopilot reviews\n- Executive Decision Packets for board, crisis, AI approval, portfolio and transformation decisions\n- Evidence classification, risk chains, decision debt, board challenges and action drafts\n- Executive Decision Defense: liability, blind spots, commitment integrity, value gates, decision SLA and kill-switches\n- Local file/directory analysis and Visual Command Center data\n\nSetup context:\n- {Only include when useful: saved preferences or source-category gaps from `cio_preflight`}\n\nGood first prompts:\n- `@The Autonomous CIO run a governed CIO autopilot review from this context.`\n- `@The Autonomous CIO build an Executive Decision Packet for this board decision.`\n- `@The Autonomous CIO stress-test this board narrative and show decision liability.`\n```\n\n## Routing\n\nIf several focused skills apply, choose the best suite first, then sequence specialist modules inside that suite. Do not expose a long list of specialist skills unless the user asks for implementation detail. Do not perform focused workflow logic here.\n\n## Skill Suites\n\nRoute through these eight suites first. Use specialist skills as internal modules unless the user explicitly asks for one.\n\n1. Executive Operating System: `autonomous-cio-operating-review`\n2. Executive Decision Intelligence: `executive-decision-packet`\n3. Board And Narrative Defense: `board-challenger`\n4. Risk, Resilience And Controls: `risk-chain-intelligence`\n5. Transformation, Portfolio And Value: `transformation-value-tracker`\n6. Industrial IT/OT, Quality And Manufacturing: `industrial-cio-operating-system`\n7. Architecture, Data, AI And Vendor: `architecture-data-security-intelligence`\n8. Organization, Operating Model And Knowledge: `organization-workforce-intelligence`\n\nSee `docs/skill-suites.md` for the specialist modules inside each suite.\n\n## Primary Workflows\n\n### autonomous-cio-operating-review\n\nUse for broad CIO reviews, `run the CIO`, monthly operating reviews, mixed enterprise context, board in 48h scenarios, crisis-to-control reviews and enterprise decision-readiness work.\n\n### executive-decision-packet\n\nUse when context must become one decision-ready artifact with decision needed, facts vs assumptions, risk chain, options, board challenge questions, Executive Decision Defense, recommendation, missing evidence and draft next steps.\n\n### autonomous-cio-orchestrator\n\nUse when the user asks a broad or ambiguous executive question and needs a chain selected before the artifact is produced.\n\n### industrial-cio-operating-system\n\nUse when the context involves regulated manufacturing, production continuity, ERP/MES/QMS/PLM, IT/OT interfaces, shopfloor dependencies, audit evidence, digital order channels or supplier/platform exposure. Produce an integrated CIO operating view, not isolated IT tickets.\n\n### executive-truth-layer\n\nUse when the main task is separating facts, assumptions, hypotheses, narratives, contradictions and missing evidence.\n\n### risk-chain-intelligence\n\nUse when isolated risks need propagation mapping across project, architecture, security, finance, vendor, compliance and operations.\n\n### decision-scenario-intelligence\n\nUse when the user asks what to approve, defer, stop, fund, re-scope or decide.\n\n### board-challenger\n\nUse before board, steering, audit, CISO/CFO or investment approval to pressure-test narrative, evidence and liability.\n\n### autonomous-action-framework\n\nUse after insights exist and the user needs action drafts, approval requests, escalation drafts, decision logs or an action ledger.\n\n### autonomous-executive-memory\n\nUse when the user asks to carry decisions, assumptions, commitments, evidence or follow-ups forward.\n\n## Source Categories\n\nThese are semantic source categories, not fixed connectors:\n\n- `calendar`: executive meetings, board dates, steering forums, review cadence.\n- `meeting_notes`: decisions, objections, commitments, risks and owner language.\n- `document_store`: board packs, architecture notes, audit evidence, policies, strategy docs.\n- `internal_messaging`: escalation texture, owner signals, blocker discussion and weak signals.\n- `delivery_system`: project status, RAID, backlog, milestones, dependencies.\n- `service_system`: incidents, outages, service health, operational risks.\n- `security_system`: findings, access, controls, exceptions, remediation.\n- `finance_system`: budget, forecast, spend, reserve, benefits and value evidence.\n- `production_system`: MES, shopfloor, machine interface, production planning and line-impact signals.\n- `quality_system`: QMS, CAPA, validation, audit, deviation and change-control evidence.\n- `ot_system`: OT network, workstation, machine-connectivity and production-control dependency signals.\n- `local_files`: user-provided files, exports, CSV, JSON, Markdown and text.\n\nUse user-provided context and local files as valid fallback when live connectors are not available.\n\n## Next Step Guidance\n\nEnd substantive outputs with exactly one concrete next action unless the user asked for no follow-up or the workflow is blocked on a required clarification. Good continuations:\n\n- ask whether to convert the output into an Executive Decision Packet\n- offer to run Executive Decision Defense\n- offer to generate the action ledger\n- offer to update proposed executive memory\n- offer to refresh Visual Command Center data from a local file\n\n## Guardrails\n\nSkills may use the Codex host LLM for semantic extraction and executive reasoning over provided context. Do not claim live connector access, automatic persistence or external execution unless a future explicit tool and user approval are available.\n"
}