{
  "status": "retrieval-completed-not-scored",
  "official_full_benchmark_run": false,
  "source_commit": "a75c964",
  "runner_sha256": "e206811d15438958805ad23fdfda49f267625444772e6538b91e30b8fdbc1624",
  "input_sha256": "1e7cf35c758aa2cfdd78c89b990e881286f1e0f563e446f23495ce130309c949",
  "runtime": {
    "distribution": "mnemosyne-memory",
    "version": "1.0.1",
    "vcs_commit": "e1ad2d0cf5ac19298547740363795c2ad25da356",
    "identity_method": "installed distribution metadata"
  },
  "verification": {
    "status": "retained-custody-and-coverage-verified",
    "manifest_sha256": "14611675f2cff3a29aa075f6fc6f726e9318284a63e96aed8e04cccebae78fec",
    "contexts": 146,
    "queries": 3671,
    "captures": 6795,
    "empty_results": 566,
    "scope": "Checks retained bytes against the supplied manifest hash and inputs. Does not independently attest execution, score answers or admit a benchmark result."
  },
  "task_counts": {
    "eventqa_65536": {
      "queries": 500,
      "empty_results": 2
    },
    "icl_banking77_5900shot_balance": {
      "queries": 100,
      "empty_results": 0
    },
    "detective_qa": {
      "queries": 71,
      "empty_results": 15
    },
    "factconsolidation_sh_6k": {
      "queries": 100,
      "empty_results": 0
    },
    "eventqa_131072": {
      "queries": 500,
      "empty_results": 11
    },
    "eventqa_full": {
      "queries": 500,
      "empty_results": 202
    },
    "longmemeval_s*": {
      "queries": 300,
      "empty_results": 186
    },
    "ruler_qa1_197K": {
      "queries": 100,
      "empty_results": 66
    },
    "ruler_qa2_421K": {
      "queries": 100,
      "empty_results": 72
    },
    "factconsolidation_mh_262k": {
      "queries": 100,
      "empty_results": 0
    },
    "factconsolidation_mh_32k": {
      "queries": 100,
      "empty_results": 1
    },
    "factconsolidation_mh_64k": {
      "queries": 100,
      "empty_results": 0
    },
    "factconsolidation_mh_6k": {
      "queries": 100,
      "empty_results": 0
    },
    "factconsolidation_sh_262k": {
      "queries": 100,
      "empty_results": 1
    },
    "factconsolidation_sh_32k": {
      "queries": 100,
      "empty_results": 0
    },
    "factconsolidation_sh_64k": {
      "queries": 100,
      "empty_results": 1
    },
    "infbench_sum_eng_shots2": {
      "queries": 100,
      "empty_results": 9
    },
    "icl_clinic150_7050shot_balance": {
      "queries": 100,
      "empty_results": 0
    },
    "icl_nlu_8296shot_balance": {
      "queries": 100,
      "empty_results": 0
    },
    "icl_trec_coarse_6600shot_balance": {
      "queries": 100,
      "empty_results": 0
    },
    "icl_trec_fine_6400shot_balance": {
      "queries": 100,
      "empty_results": 0
    },
    "recsys_redial_full": {
      "queries": 200,
      "empty_results": 0
    }
  },
  "limitations": [
    "Mnemosyne CLI integration, not upstream full agent runner",
    "Retrieval only; no generated answers or quality scores",
    "CLI wall times include process startup; not engine-only latency",
    "Default retrieval settings; not resource-matched to official BM25.",
    "Empty retrieval results are not answer-quality scores.",
    "Retained-artifact verification is not independent execution attestation."
  ],
  "download_scope": "Receipt only; raw capture and query artifacts retained locally."
}
