Update to Aivana CIO Decision Twin

{
  "name": "autonomy-stress-test",
  "description": "Stress-test proposed autonomous CIO preparation against evidence changes, approval gaps and human-control failures. Use when the user needs autonomy stress test for CIO decision support.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 136
    }
  ],
  "skill_md_contents": "---\nname: autonomy-stress-test\ndescription: Stress-test proposed autonomous CIO preparation against evidence changes, approval gaps and human-control failures. Use when the user needs autonomy stress test for CIO decision support.\n---\n\n# Autonomy Stress Test\n\n## Mission\n\nFind where governed autonomy would fail before any action is treated as approval-ready.\n\n## Inputs\n\nAccept action ledgers, autonomy gates, decision packets, risk registers, approval notes, evidence gaps and policy constraints.\n\n## Workflow\n\n1. Identify proposed autonomous or semi-autonomous preparation steps.\n2. Test failure modes such as stale evidence, missing approver, urgency bypass and human-only domain confusion.\n3. Classify stress posture as standard guardrails, tight guardrails or fail-close.\n4. Define required safeguards before autonomy can increase.\n5. Keep the result as a decision-support artifact only.\n\n## Output Format\n\n- Executive Summary\n- Stress Score\n- Stress Posture\n- Stressors\n- Failure Modes\n- Required Safeguards\n- Human-Control Notes\n\n## Guardrails\n\nDo not authorize execution. Use the stress test to reduce autonomy where evidence or approval is weak.\n"
}