{
  "id": 13257254,
  "title": "Auditable agents: turn the answer into a claim you can check",
  "url": "https://urgent.news/2026/10/09/auditable-agents-turn-the-answer-into-a-claim-you-can-check",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-09T23:32:01.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/bzdvdn/auditable-agents-turn-the-answer-into-a-claim-you-can-check-33i0"
  },
  "original_language": "en",
  "account": "The article discusses the importance of making the underlying state of an AI agent's answer inspectable and reproducible. The author proposes treating the agent's output as a \"typed artifact\" within a versioned context, rather than a message in a bag. By linking each derived answer to its inputs as \"typed edges,\" the entire reasoning process becomes a graph that can be walked to determine the origin of the answer. This approach allows for easy identification of the 'why' behind the answer and for checking if the same result would be produced in a repeat run by comparing context hashes. The author presents a Python example using the Reactifact framework to demonstrate how to implement this reproducible answer approach, including the creation of artifacts, versioned context commits, and queryable provenance edges.",
  "summary": "A model tells you \"Q2 cloud spend was $45,000 — a 12.5% variance over budget.\" Now two uncomfortable questions: Why? Which spreadsheet cell, which policy clause, which calculation? Again? If you run it tomorrow, do you get the same answer — and can you prove it's the same? Most agent frameworks can't answer either cleanly. The run is a stream of messages; the \"reasoning\" is prose in a log; the…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}