{
  "schema": "vedokrok.public-item.v1",
  "release_id": "MHC-RPUB-20260920-75ad787a",
  "url": "/knowledge/capture-semantic-boundaries-not-only-raw-logs",
  "id": "MHC-D-RESEARCH-1159",
  "version": "0.1.0",
  "title": "Capture semantic boundaries, not only raw logs",
  "summary": "A log can be complete and still omit the boundary where meaning changed.",
  "kind": "checklist",
  "body": "Record the inputs and outputs at the model and tool boundaries that matter to the task: model request and response identifiers, parsed arguments, tool results, policy decisions and the state carried into the next step. This creates an execution envelope that can replay the bad decision against repaired code.",
  "limits": [
    "Do not turn observability into indiscriminate data retention. Minimize personal data, secrets and irrelevant payloads."
  ],
  "topics": [
    "union-ai-runtime-verification-2026"
  ],
  "intents": [],
  "source_ids": [
    "RS-F0FB27D3E1B6C4A1"
  ],
  "evidence": [
    {
      "claim": "Recording semantic model and tool boundaries can preserve the decision context needed to replay a production failure against repaired enforcement code.",
      "source_id": "RS-F0FB27D3E1B6C4A1",
      "role": "supports",
      "note": "Capturing more data is not automatically useful or safe; retain only the boundaries needed for diagnosis while respecting privacy and secret-handling requirements.",
      "locator": "4:51-13:14, recover the run, record semantic boundaries, replay the bad decision and retain the execution envelope"
    }
  ],
  "use_when": [
    "A production agent failure disappears when developers try to reproduce it locally."
  ],
  "avoid_when": [
    "Do not turn observability into indiscriminate data retention. Minimize personal data, secrets and irrelevant payloads."
  ],
  "example": "A trading assistant interpreted dollars as shares. The useful trace preserves the model's action, parsed amount, tool contract and validation result at the order boundary.",
  "check": "A developer can replay the consequential boundary without inventing missing inputs or re-prompting the model for a similar decision.",
  "checklist": [
    "Capture the model decision that selected the action.",
    "Capture the exact parsed tool arguments and returned result.",
    "Record policy or validation decisions around the call.",
    "Keep enough state to replay the boundary without storing unrelated sensitive data."
  ],
  "sources": [
    {
      "id": "RS-F0FB27D3E1B6C4A1",
      "title": "Your Agent Failed in Prod. Good Luck Reproducing It.",
      "url": "https://ai.engineer/talks/Lc8zRh9muoY-your-agent-failed-in-prod-good-luck"
    }
  ],
  "relations": [
    {
      "from": "MHC-D-RESEARCH-1159",
      "to": "MHC-D-RESEARCH-0958",
      "type": "useful_with",
      "url": "/knowledge/track-tool-errors-tool-calls-runtime-and-tokens-as-agent-diagnostics"
    }
  ],
  "collections": [
    {
      "id": "RC-ACD6385FB70207D3",
      "title": "Operate AI agents as systems you can replay, verify and constrain",
      "url": "/collections/operate-ai-agents-as-systems-you-can-replay-verify-and-constrain"
    }
  ]
}