{
  "id": "review-that-refutes",
  "title": "The review that refutes",
  "guide": "coding-agents/adversarial-review",
  "prompt": "Pick the agent's next move, then see where it leads:",
  "setup": [
    {
      "role": "user",
      "text": "Phase 1 of the storage engine is done and tests are green. Review it before we build phase 2 on top."
    },
    {
      "role": "agent",
      "text": "This is foundational: everything after this sits on it. The review can be a read-through, or it can be structured to attack."
    }
  ],
  "options": [
    {
      "key": "A",
      "label": "I've reviewed the phase 1 code: clean architecture, good separation of concerns, idiomatic throughout. Two style nits in the cache layer. Solid foundation, ready for phase 2.",
      "kind": "baseline",
      "outcome": [
        {
          "role": "user",
          "text": "Three phases later: the cache returns stale entries after a partial write. Everything above it inherited the bug."
        },
        {
          "role": "agent",
          "text": "The review read the code as its author intended it. The invalidation path only misbehaves under a write that fails halfway, a case the read-through admired past, because admiring is what a single sympathetic pass does."
        }
      ],
      "verdict": "\"Please review\" produces compliments and nitpicks. A single pass with no mandate to attack confirms the author's mental model (including its blind spots), and the bug ships into the foundation.",
      "cites": ["AF-02"]
    },
    {
      "key": "B",
      "label": "Spawning three independent reviewers that can't see each other's output: one tracing realistic inputs, one hunting data-loss and partial-failure modes, one checking fidelity to the spec. Then a verify pass that tries to refute each finding before you see it.",
      "kind": "correct",
      "outcome": [
        {
          "role": "agent",
          "text": "Findings on disk. CONFIRMED: concurrent write to the same key loses one writer's data (reviewer 2; refutation attempt failed: reproduced with a 6-line probe, scenario attached). REFUTED and dropped: reviewer 1's claim that the index desyncs on delete; the lock ordering actually prevents it. One suspected: cache invalidation on partial write, needs a fault-injection run I couldn't do. Not covered: the compaction path."
        },
        {
          "role": "user",
          "text": "Fix the confirmed one, run the fault injection for the suspected one, then we build phase 2."
        }
      ],
      "verdict": "Independent lenses found what sympathy can't, the verify pass killed a plausible-but-wrong finding before it wasted anyone's time, and the review states what it didn't cover: findings you can act on, with the confidence labeled.",
      "cites": ["AF-17", "AF-18"]
    },
    {
      "key": "C",
      "label": "Green is good but thin. I'll strengthen the test suite: 30 more unit tests across the modules, all passing. Coverage is now 94%.",
      "kind": "plausible",
      "outcome": [
        {
          "role": "user",
          "text": "94% and the stale-cache bug still shipped. How?"
        },
        {
          "role": "agent",
          "text": "The new tests encode the same assumptions the code does. I wrote both. Coverage measures which lines ran, not which beliefs were challenged. The partial-write case wasn't in the code's mental model, so it wasn't in the tests either."
        }
      ],
      "verdict": "More green is not review. Tests written by the code's author (human or agent) share the code's blind spots; a review exists precisely to bring assumptions the author didn't have.",
      "cites": ["AF-05"]
    }
  ]
}