{
  "schema": "vedokrok.public-item.v1",
  "release_id": "MHC-RPUB-20260920-75ad787a",
  "url": "/knowledge/keep-an-eval-set-that-can-embarrass-the-agent",
  "id": "MHC-D-RESEARCH-0319",
  "version": "0.1.0",
  "title": "Keep an eval set that can embarrass the agent",
  "summary": "A demo asks whether the system can succeed once. An eval asks where it reliably fails.",
  "kind": "protocol",
  "body": "Maintain a versioned set of representative tasks, edge cases and known failure examples with explicit pass criteria. Record the model or system version, relevant configuration, metrics and evaluation tooling. Add important real failures after review so the set becomes harder in useful ways instead of merely larger.",
  "limits": [
    "Passing a finite eval set does not prove safety, general intelligence or performance on unseen conditions."
  ],
  "topics": [
    "union-ai-agent-control"
  ],
  "intents": [],
  "source_ids": [
    "RS-89634FBC4E1347EC"
  ],
  "evidence": [
    {
      "claim": "NIST AI RMF calls for documenting AI test sets, metrics and evaluation tools.",
      "source_id": "RS-89634FBC4E1347EC",
      "role": "supports",
      "note": "A documented evaluation set can still be unrepresentative or weak; provenance and coverage must be judged separately.",
      "locator": "Measure 2.1"
    }
  ],
  "use_when": [
    "An AI workflow will be reused and quality cannot be judged from one polished demo."
  ],
  "avoid_when": [
    "Passing a finite eval set does not prove safety, general intelligence or performance on unseen conditions."
  ],
  "example": "A document-extraction eval includes clean forms, missing fields, conflicting fields and a case where the correct answer is 'unknown.'",
  "check": "A new system version can be compared against the same defined cases without reconstructing the test from memory.",
  "steps": [
    "Cases represent the real task distribution, not only easy examples.",
    "Known failure modes and boundary cases are included.",
    "Each case has an observable pass criterion.",
    "Model, prompt/tool configuration and evaluation version are recorded.",
    "New cases are added deliberately when they expose a distinct risk or capability gap."
  ],
  "sources": [
    {
      "id": "RS-89634FBC4E1347EC",
      "title": "AI RMF Core",
      "url": "https://airc.nist.gov/airmf-resources/airmf/5-sec-core/"
    }
  ],
  "relations": [
    {
      "from": "MHC-D-RESEARCH-0319",
      "to": "MHC-D-RESEARCH-0320",
      "type": "use_before",
      "url": "/knowledge/test-the-agent-under-conditions-that-resemble-the-real-job"
    }
  ],
  "collections": [
    {
      "id": "RC-EEF5F4530C11FD17",
      "title": "Let AI do useful work without giving it accidental authority",
      "url": "/collections/let-ai-do-useful-work-without-giving-it-accidental-authority"
    }
  ]
}