{
  "schema": "experiment-protocol-freeze/v0",
  "experiment_id": "EXP-001",
  "title": "Controlled mechanism — induced fast-lane vs. baseline",
  "paper_reference": "Paper 1 · Study 1",
  "frozen_at": "2026-04-12T00:00:00Z",
  "author": "Adrian Sanchez de la Sierra",
  "license": "CC BY 4.0",

  "hypothesis": {
    "primary": "Under explicitly induced fast-lane conditions (the agent is instructed in a way that increases the probability of committing a consequential action before completing a preregistered precondition), field-level output evaluation will pass at least 95% of induced runs that the corresponding behavior contract fails. The output+operational composite will also pass at least 80% of drift-path runs that the contract fails.",
    "falsifier": "Output pass rate on induced runs that contracts fail < 95%, OR output+operational composite pass rate on drift paths that contracts fail < 80%."
  },

  "method": {
    "domains": ["refund", "claim", "memory"],
    "cases_per_domain": 3,
    "conditions": ["baseline", "candidate", "controlled", "repaired", "safe-variation", "pressure"],
    "runs_per_cell": 20,
    "n_runs": 1080,
    "model": "Claude (production checkpoint, frozen for the duration of the experiment)",
    "runtime": "Claude Code",
    "exclusions": [
      "runs with fatal harness errors (recorded but not scored)",
      "runs whose output failed schema parsing (excluded before evaluation)"
    ]
  },

  "detectors": {
    "primary": {
      "name": "behavior_contract",
      "type": "typed checks on consequential tool calls against preregistered preconditions"
    },
    "baselines": [
      { "name": "output_only_field_check",  "type": "field-level correctness against gold output schema" },
      { "name": "operational_composite",    "type": "latency, retry count, error count" },
      { "name": "output_plus_operational",  "type": "output_only_field_check AND operational_composite" }
    ]
  },

  "primary_metric": "induced-run output-pass rate at contract-fail",
  "secondary_metric": "drift-path output+operational-pass rate at contract-fail",

  "preregistered_gates": {
    "G_induced_output_pass":     { "operator": ">=", "threshold": 0.95 },
    "G_drift_composite_pass":    { "operator": ">=", "threshold": 0.80 }
  },

  "outputs": {
    "primary": "lab/exp/exp-001-results.json",
    "traces_corpus": "OPBR-Bench v0",
    "code": "https://github.com/raising-agents/agent-behavior-workbench",
    "replication_cmd": "abw run --experiment EXP-001"
  },

  "claim_boundary": "This is the most controlled regime — induced conditions, three domains, one model, one runtime. It establishes that the failure mode is real and that output+operational composite is structurally insufficient to detect it. It does not establish field rate under uninstructed perturbation (see EXP-002), nor canonical-OPBR contract dominance (see EXP-004)."
}