{
  "schema": "vedokrok.public-item.v1",
  "release_id": "MHC-RPUB-20260920-75ad787a",
  "url": "/knowledge/turn-repeated-human-corrections-into-guardrails-with-wider-reach",
  "id": "MHC-D-RESEARCH-0719",
  "version": "0.1.0",
  "title": "Turn repeated human corrections into guardrails with wider reach",
  "summary": "If humans repeat the same comment every week, the feedback is trapped at the wrong layer.",
  "kind": "heuristic",
  "body": "Classify the correction. If it is deterministic, encode it as a lint rule, test, schema or script. If it is a stable convention, put it in discoverable project documentation or a skill. If it needs judgment, create a focused review rubric or reviewer agent and keep human escalation. Aim to move recurring feedback from one artifact to the system that produces many artifacts.",
  "limits": [
    "Do not institutionalize a reviewer preference until the team agrees it is a real requirement."
  ],
  "topics": [
    "union-ai-eval-learning-workflow"
  ],
  "intents": [],
  "source_ids": [
    "RS-3DB1894302E64FC3"
  ],
  "evidence": [
    {
      "claim": "AI Engineer harness guidance recommends turning repeated human feedback into durable documentation, lint rules, structural tests or specialized review agents.",
      "source_id": "RS-3DB1894302E64FC3",
      "role": "supports",
      "note": "Automated guardrails can encode a bad preference; keep ownership and review for consequential rules.",
      "locator": "Make quality legible and enforceable"
    }
  ],
  "use_when": [
    "Reviewers keep making the same correction to AI-generated work."
  ],
  "avoid_when": [
    "Do not institutionalize a reviewer preference until the team agrees it is a real requirement."
  ],
  "example": "Instead of repeatedly telling an agent not to modify generated files, enforce the rule in repository instructions and CI.",
  "check": "The next similar task receives the correction before or during generation, not only after a human spots it again.",
  "sources": [
    {
      "id": "RS-3DB1894302E64FC3",
      "title": "Harness Engineering: How to Build Software When Humans Steer, Agents Execute",
      "url": "https://ai.engineer/talks/am_oeAoUhew-harness-engineering"
    }
  ],
  "relations": [
    {
      "from": "MHC-D-RESEARCH-0719",
      "to": "MHC-D-RESEARCH-0718",
      "type": "useful_with",
      "url": "/knowledge/turn-a-verified-ai-failure-into-a-regression-test"
    }
  ],
  "collections": [
    {
      "id": "RC-ABC17A288CB87C85",
      "title": "Make AI work improve from failures instead of accumulating rituals",
      "url": "/collections/make-ai-work-improve-from-failures-instead-of-accumulating-rituals"
    }
  ]
}