{
  "dataset": "glossary",
  "record": {
    "id": "action-audit",
    "term": "Action Audit Log",
    "category": "optimization",
    "short_def": "A visible, timestamped record of the actions an agent has taken.",
    "long_def": "The 'what did happen' half of agent trust (intent preview is the 'what will happen' half). An auditable trail lets a principal verify and, if needed, undo what an agent did.",
    "see_also": [
      "intent-preview",
      "agent-experience"
    ],
    "etymology_origin": {
      "value": null,
      "verify_status": "verify-against-primary-at-build",
      "source_hint": "https://biilmann.blog/articles/introducing-ax/ — 'action audit log' is an Agent Experience (AX) trust pattern; audit logging itself is a general security concept with no single coiner for the agent sense"
    },
    "related_to": [
      "intent-preview",
      "agent-experience"
    ],
    "contrast_with": "Unlike intent preview, which asks for approval before an action, an action audit log is the after-the-fact record — the 'what did happen' half — that lets a principal verify or undo an agent's work.",
    "example": "After an agent completes a task, its action audit log shows a timestamped trail of every call it made, letting the principal verify and, if needed, undo each one.",
    "source": "https://biilmann.blog/articles/introducing-ax/",
    "status": "emerging",
    "why_it_matters": "An action audit log is the accountability half of agent trust; without it, a principal cannot verify or reverse what an autonomous agent did on their behalf.",
    "sameAs": [],
    "bridge_entity": "agent-readiness",
    "last_verified": "2026-06-15",
    "md_twin": "/glossary/action-audit.md"
  }
}