{
  "argument": "Machine-speed containment can reduce loss only when false-positive effects cannot exceed the threat being controlled.",
  "canonicalUrl": "https://xn--mwe.com/assurance-case/claims/autonomous-defense-is-bounded-and-reversible/",
  "claim": "Autonomous defenders act only through pre-authorized internal actions whose physical consequence, confidence threshold, rollback, evidence, and safe-state behavior have been tested.",
  "defeaters": [
    "unbounded tool access",
    "unknown physical consequence",
    "model drift",
    "poisoned telemetry",
    "rollback failure",
    "conflicting agent state"
  ],
  "evidenceRequirements": [
    "allowed-action policy",
    "causal dependency model",
    "threshold and uncertainty records",
    "rollback tests",
    "false-positive consequence analysis",
    "model and data provenance",
    "decision receipts"
  ],
  "hazards": [
    "self-inflicted outage",
    "unsafe OT isolation",
    "cascading agent failure",
    "evidence loss"
  ],
  "id": "AC-K05-C05",
  "name": "Autonomous defense is bounded, causal, and reversible",
  "recoveryObjectives": [
    "stop action on uncertainty",
    "restore last known-good state",
    "retain incident chronology",
    "escalate unresolved state"
  ],
  "releaseId": "K12-2026-08-16",
  "siteTailoring": [
    "action tier",
    "protected process",
    "latency requirement",
    "operator authority",
    "maintenance state"
  ],
  "slug": "autonomous-defense-is-bounded-and-reversible",
  "sourceIds": [
    "SRC-NIST-800-82R3",
    "SRC-NIST-800-207",
    "SRC-WH-GOLD-EAGLE-2026"
  ]
}