Skip to content
AIGP

RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence — 17. Replay Semantics

AIGP Specification › RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence › 17. Replay Semantics

← 16. D-DNA for Governed Autonomy · Section index · 18. Force-Capable Systems →

17. Replay Semantics

AIGP must support replay of autonomous legitimacy.

Replay must answer:

Was the system authorized?
What governance intent applied?
Was the action inside the operational envelope?
Which symbolic constraint governed the action?
Was the enforcement state active?
Was human authority required?
Was human authority present?
Was uncertainty within threshold?
Was the action permitted, denied, modified, or escalated?
Was the consequence observed?
Was the D-DNA chain intact?

Replay modes:

replay_modes:
  governance_replay:
    purpose: reconstruct authority, intent, constraint, decision, and evidence


  operational_replay:
    purpose: reconstruct system state, telemetry, environment, and action


  accountability_replay:
    purpose: reconstruct human authority, escalation, override, harm, and consequence

Replay must not depend on natural-language explanation.

The source of truth is:

signed D-DNA event chain
canonical symbolic constraint
operational envelope version
runtime context
evidence hashes

← 16. D-DNA for Governed Autonomy · Section index · 18. Force-Capable Systems →