{
  "protocol": {
    "episodes": 120,
    "memory_size": 24,
    "seed": 42,
    "baseline": "long-context on the same mini-suite",
    "fidelity": "mechanism reproduction on deterministic benchmark mini-suites"
  },
  "baseline": {
    "method": "long-context",
    "joint_success": 1.0,
    "average_cost": 64.5
  },
  "methods": {
    "mrkl": {
      "benchmark": "scalemcp-mini",
      "joint_success": 1.0,
      "average_cost": 1.25,
      "router_calls": 360,
      "symbolic_expert_calls": 170
    },
    "hugginggpt": {
      "benchmark": "planbench-mini",
      "joint_success": 1.0,
      "average_cost": 2.350000000000002,
      "model_matches": 360,
      "dependency_edges": 240
    },
    "generative-agents": {
      "benchmark": "evomem-mini",
      "joint_success": 1.0,
      "average_cost": 1.7900000000000034,
      "memories_retrieved": 354,
      "reflection_syntheses": 30
    },
    "memgpt": {
      "benchmark": "evomem-mini",
      "joint_success": 1.0,
      "average_cost": 0.9200000000000021,
      "archival_writes": 108,
      "page_ins": 96,
      "interrupts": 108
    }
  },
  "interpretation": "The deterministic suites test state transitions and cost under a shared protocol. Success=1.0 is not an original-paper benchmark score.",
  "schema_version": 2,
  "manifest_ref": "experiments:p0-missing-agent-mini-suites-seed42",
  "evaluation_protocol": {
    "tier": "l1_mechanism",
    "seeds": [
      42
    ],
    "formal_comparison": false,
    "claim_policy": "single/few-seed smoke result; do not claim a stable improvement"
  },
  "provenance": {
    "artifact_path": "docs/experiments/p0-missing-agent-mini-suites-seed42.json",
    "historical_migration": "historical-metrics-v2-2026-08-09",
    "original_code_commit": "not recorded",
    "dataset_fingerprint": "not recorded in historical artifact"
  }
}