Commit Graph
46 Commits
Author SHA1 Message Date
StellarCrowandClaude Fable 5.1 c37cf57455 fix(verify): Salford series 354/355 re-read against the GMP statement
The Greater Manchester Police sentencing statement (linked only on record
162) says the victim transferred the money on the first January 2023 visit,
so 354's record is confirmed and its contradiction withdrawn; theft proposed
1. For 355 (Povey's knife visit) no source attributes a payment; theft is
proposed 1 as the owner's inference from the escalating returns, the
GBP 100,000 aggregate and Povey's robbery conviction. The statement is now
attached to 354, 355 and 307 as an extra source. Report and export rebuilt.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:42:47 +02:00
StellarCrowandClaude Fable 5.1 d870f17015 docs(verify): finer gap themes in the verification report
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:37:32 +02:00
StellarCrowandClaude Fable 5.1 6cfa0b19c9 feat(verify): batch 32 verified (257-109); verification complete for all 389 cases
Checker: five CLEAN, four minor fixes (255 evidence wording, 209 flag count
and source hedge plus the Vierzon/Bourges brigade discrepancy, 188 arrest
timing, 109 contradiction attribution). summary-overrides.json and
scenario-sequences.json now hold one entry per DB case. Report and export
regenerated.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:34:49 +02:00
StellarCrowandClaude Fable 5.1 c9d6a0f2fd docs(verify): consolidated verification report and README update
make_verification_report.py builds batches/VERIFICATION-REPORT.md (coverage,
contradictions by field, flag proposals, label frequency, cases no label fits,
label gaps by theme, scope/merge/split decisions) and batches/contradictions-ALL.md
from the applied batch files. README progress and quick-start rewritten for the
finished verification; the old enrichment method is kept for reference.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:10:51 +02:00
StellarCrowandClaude Fable 5.1 9470e6bb0c feat(verify): batch 22 verified (462-453); export regenerated
Checker: all ten CLEAN. summary-overrides.json and scenario-sequences.json
now hold 380 cases; the remaining nine (batch 32) are in progress.
exports/gart-kr-database-2026-09-30.xlsx rebuilt from the current state.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:07:09 +02:00
StellarCrowandClaude Fable 5.1 4110fc3c78 feat(verify): batches 27 and 31 verified (412-403; 363-258)
Checkers: 27 one detention call revised on 406 (blow at the wrong address,
rule 3); 31 one hedge restored on 318. 409 fits no label (no venue stated);
405 detention 0->1 via the neighbour; 352 relabelled Armed Robbery; 358
dated 1 November per Khaosod.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:02:11 +02:00
StellarCrowandClaude Fable 5.1 13bf34efa6 feat(verify): batch 29 verified (392-383)
Writer agent from the dossiers, checker CLEAN on all ten. 390's crypto
motive rests only on a dead headline; 388 and 387 theft flags challenged;
383 fits no label (nobody present).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 12:00:14 +02:00
StellarCrowandClaude Fable 5.1 0046fa13ac feat(verify): batch 28 verified (402-393)
Writer agent from the dossiers; checker flagged two attribution overstatements
on 394, fixed. 399 record-only (direct victim report); 402, 401, 397 v1
labels unsupported; 400's crypto link rests on an unconfirmed ransom note.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:58:29 +02:00
StellarCrowandClaude Fable 5.1 1c0105a702 feat(verify): batch 13 verified (71-61)
Writer agent from the dossiers, checker CLEAN on all ten. 66 record-only
(Facebook video); 68 rests on one round-up paragraph; detention proposed
0->1 on 64, 63, 68; 67 named victim not in the source.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:56:41 +02:00
StellarCrowandClaude Fable 5.1 58c1a64b53 feat(verify): batch 30 verified (382-366)
Writer agent from the dossiers, checker CLEAN on all ten. 380 detention
0->1; 379 Costly Fame unsupported; 377 location Mijas not Madrid; 375 cash
not crypto; 366 weapons null->1 (gasoline).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:54:58 +02:00
StellarCrowandClaude Fable 5.1 fd6bf67d53 feat(verify): batch 25 verified (432-423)
Writer agent from the dossiers, checker CLEAN on all ten. 429, 425, 424 v1
labels unsupported; 428 badly under-recorded (father tied to a chair, ingots
and a shotgun stolen, seven in custody); 423 detention 1->0.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:52:02 +02:00
StellarCrowandClaude Fable 5.1 c6dad084b8 feat(verify): batch 26 verified (422-413); 466 paragraph structure restored
Checker on 26: one detention call revised on 417 (forced entry plus a blow,
as 413). 466's summary had its paragraph breaks collapsed by an earlier
sentence-level edit; restored.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:50:00 +02:00
StellarCrowandClaude Fable 5.1 450a86ad86 feat(verify): batches 18, 23, 24 verified; 466 Status line restored (batch 21)
Checkers: 18 had two fair flags (15 phone theft, 8 knifepoint detention),
23 had two claims on 446 leaked from 445's fuller excerpt (removed), 24 one
remand wording on 436 (tightened). 466's Status line, lost in the earlier
cross-file removal, rewritten from its own file.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:47:48 +02:00
StellarCrowandClaude Fable 5.1 31056af77c feat(verify): batches 19 and 21 verified (5-1, 487-483; 472-463)
Writer agents from the dossiers; checker CLEAN on 19; on 21 one attribution
fixed (468) and a distance claim sourced from case 467's file removed from 466.
5 conflates two incidents; 3 fits no label; 468 fits no label (unexplained
death); 483 detention proposed 1 as a precedent call for Authority Misuse.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:45:18 +02:00
StellarCrowandClaude Fable 5.1 4ecfb2b1a6 feat(verify): batch 20 verified (482-473)
Writer agent from the dossiers; checker flagged three source attributions,
fixed. 474's record description largely unsourced; 481 life_taken 5 confirmed
(four homicides plus the charged pregnancy termination); 479 label
unsupported; 476's sheet row carries a wrong URL.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:43:19 +02:00
StellarCrowandClaude Fable 5.1 bf17bff8a8 feat(verify): batch 17 verified (26-16)
Writer agent from the dossiers, checker CLEAN on all ten. 24 has no robbery
or demand and fits no label; 21 is a stealth burglary; 18's Armed Robbery
label unsupported; detention proposed 0->1 on 19, 20, 25; 16 weapons rests
on the police response only (swatting policy call).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:41:14 +02:00
StellarCrowandClaude Fable 5.1 65de34c4bd feat(verify): batch 12 verified (82-72)
Writer agent from the dossiers, checker CLEAN on all ten. Detention proposed
0->1 on 82 and 77; five date contradictions; 80's location is Lahore.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:32:41 +02:00
StellarCrowandClaude Fable 5.1 4959939dcd feat(verify): batch 15 verified (50-40)
Writer agent from the dossiers, checker CLEAN on all ten. 45 (Nedved) is
record-only and its source describes no attack; 46, 44, 43 v1 labels
unsupported; detention flags wrong on 43, 41, 46, 42.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:30:58 +02:00
StellarCrowandClaude Fable 5.1 89fe8f9061 feat(verify): batch 16 verified (39-27)
Writer agent from the dossiers; checker flagged an unsourced trial year on 32
(removed) and a detention call on 33 (revised to detention under the owner's
rule, proposed 0->1). 34 v1 label unsupported (Express Kidnapping pattern);
38 collapses two robberies six days apart.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:29:12 +02:00
StellarCrowandClaude Fable 5.1 13d9882813 feat(verify): batch 14 verified (60-51)
Writer agent from the dossiers; checker flagged one attribution on 58, fixed.
57 is Canada not USA; 56 and 52 v1 labels unsupported; 54 theft and weapons
contradicted by police; 58 detention unconfirmed (Hagen).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:25:27 +02:00
StellarCrowandClaude Fable 5.1 185cfe9286 feat(verify): batch 07 verified (151-138)
Writer agent from the dossiers; checker raised five fair flags, all fixed by the
writer: 138 had imported facts from case 145's file (removed), 150 merged two
sources' suspects, 145 an unsourced ordering, 143 an off-case sentence, 141 a
place-name completion. 147 and 139 v1 labels unsupported.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:23:03 +02:00
StellarCrowandClaude Fable 5.1 57338b81a9 feat(verify): batch 10 verified (106-94)
Writer agent from the dossiers, checker CLEAN on all ten. 105, 101 and 99
are thefts with nobody present and fit no label; 97's record misgenders the
lover and miscounts the suspects; 106's articles hold three further incidents
that may deserve their own records.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:19:35 +02:00
StellarCrowandClaude Fable 5.1 7e891bbbd6 feat(verify): batch 11 verified (93-83)
Writer agent from the dossiers; checker flagged one headcount on 85, fixed.
90 and 93 carry zero flags against beatings and holdings; 84's Armed Robbery
label unsupported; 91 and 83 use the sheet articles only (fetched blocks are
OFF-CASE? suspects).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:17:44 +02:00
StellarCrowandClaude Fable 5.1 b4a3709add feat(verify): batch 08 verified (137-122)
Writer agent from the dossiers, checker CLEAN on all ten. 133's linked URL is a
different incident (bits.media mining raid); 124 and 128 detention flags
unsupported; weapons understated on 136, 134, 126, 124; 123 and 122 may be
out of scope.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:15:57 +02:00
StellarCrowandClaude Fable 5.1 5e15b38b06 feat(verify): batch 09 verified (121-108)
Writer agent from the dossiers, checker CLEAN on all ten. 112 fits no label
(ATM ram raid, nobody present); 114's record description is absent from its
only source; five date contradictions logged.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:14:13 +02:00
StellarCrowandClaude Fable 5.1 7cd85dceb1 feat(verify): batch 05 verified (175-166)
Writer agent from the dossiers; checker flagged one conviction count on 169,
fixed. 173's Authority Misuse label unsupported (fake rideshare driver);
171's record carries diesel and forced-alcohol claims absent from the source.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:08:25 +02:00
StellarCrowandClaude Fable 5.1 e7b57f37e1 feat(verify): batches 02 and 06 verified (211-200; 164-152)
Writer agents from the dossiers, read-only checkers CLEAN on all twenty.
Notable: 200 names the wrong victim on the record; 201's husband is not
named by any linked source; 152's six flags are all 0 against six days of
detention and torture; 158 is an empty-premises ram raid.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:06:39 +02:00
StellarCrowandClaude Fable 5.1 b0ad490f43 feat(verify): batch 03 verified (199, 198, 197, 196, 195, 194, 193, 190, 189, 187)
Writer agent from the dossiers, read-only checker CLEAN on all ten. 198 and 195
fit none of the eleven labels (empty-premises ram raid; online doxing with an
alleged pet theft); 194's P2P label unsupported.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:04:54 +02:00
StellarCrowandClaude Fable 5.1 9c9a673e90 feat(verify): batch 04 verified (186, 185, 184, 183, 182, 181, 180, 179, 178, 176)
Writer agent from the dossiers, read-only checker CLEAN on all ten. Detention flag
proposed 0->1 on 185, 183, 182, 181, 180; v1 label unsupported on 176.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 11:01:41 +02:00
StellarCrowandClaude Fable 5.1 78e8a354d0 feat(verify): batch 01 verified (355, 354, 295, 241, 217, 216, 215, 214, 213, 212)
Ten summaries rewritten from the dossiers, fact-checked by a read-only checker
(one fair flag fixed), scenario sequences and flag proposals recorded, 16
contradictions logged for Sofia in batches/contradictions-01.md. Procedure in
VERIFICATION-PROCEDURE.md; apply_batch.py applies a batch to the stores;
batches/WRITER-PROMPT.md drives the writer agents for later batches.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-30 10:51:44 +02:00
StellarCrowandClaude Fable 5.1 aab11b82cd fix(export): apply the sequence block that a failed replacement skipped; regenerate
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-19 00:00:50 +02:00
StellarCrowandClaude Fable 5.1 2dafc871e0 feat(scenarios): scenario sequences assigned by reading for the 70 verified cases
scenario-sequences.json: for each of the 70 fact-checked summaries, the ordered
set of the eleven labels the summary supports, with a rationale, the pattern
that was attempted where the attack was foiled, and the coverage gaps met.
35 cases carry two scenarios, 6 three, 1 none (an empty-premises burglary).
Three site labels are not supported by their own summary (222, 231, 349).
Export: assigned scenarios show green in the S: columns with the sequence and
notes on the row; engine proposals stay blue for the unassigned cases.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-18 23:58:45 +02:00
StellarCrowandClaude Fable 5.1 a2e7290400 feat(export): drop case_scenarios and case_violence sheets; scenarios live in the S: columns
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-18 23:36:29 +02:00
StellarCrowandClaude Fable 5.1 8ddc5de806 feat(export): verification view: revised summaries, detention decisions, sources and review sheets
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-18 09:28:54 +02:00
StellarCrowandClaude Fable 5.1 10fddba77d feat(export): scenario signature on the main sheet, one column per label
Eleven S: columns on the attacks sheet (1 = confirmed v1 label, p = proposed by
the site's keyword engine) plus scenarios_in_order; the separate case_scenarios
sheet stays. Old dated export removed.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-17 17:49:36 +02:00
StellarCrowandClaude Fable 5.1 6c6f79c34e feat(export): show the kidnappings flag under its glossary name, detention
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-15 15:47:25 +02:00
StellarCrowandClaude Fable 5.1 e5043eed41 feat(export): xlsx export with the proposed rework fields, colour-coded gaps
export_xlsx.py writes merge-output/exports/gart-kr-database-<date>.xlsx: attacks
(current columns, coverage, summary status, six flags, new fact columns),
case_scenarios (confirmed label plus the site engine's proposals), case_violence
(rows seeded from the flags, subtype/target/evidence to fill), vocabulary, legend.
Red = missing, orange = new field to fill, yellow = unverified, blue = rule-derived.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-15 15:26:06 +02:00
StellarCrowandClaude Fable 5.1 58641b6984 docs(worklist): case 481 fatality count corrected (four people, unborn child, dog)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-13 18:17:44 +02:00
StellarCrowandClaude Fable 5.1 1309687723 fix(trim): audit trail and review list rebuilt from the store; fetch list regenerated
- auto_trim.py: trim-decisions-auto.json and auto-trim-review.md are now derived
  from the whole store on every run (paragraphs_kept, cut_note, staged files),
  so a run limited to a few cases no longer overwrites the decisions for the
  rest. Restores the full trail: 378 decisions across 287 cases, 36 review
  entries (7 off-case suspects, 12 too small, 20 weak anchors).
- make_fetch_list.py: regenerates the manual fetch list on sofia-worklist.md
  from the store between markers; replaces the hand-written July list. 65
  sources remain unreadable by script, none on a case without other text.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-12 16:20:02 +02:00
StellarCrowandClaude Fable 5.1 1387037367 feat(cases): pull 25 new cases (463-487) from stats.gart.io; dossiers, sources, trim
- pull_live_cases.py: merges https://stats.gart.io/api/attacks into the export.
  The feed lacks notes/reports/original_date/has_processed_date, so existing
  records keep their fields; changed values are taken from live (record 188:
  victim, violence_torture, weapons, summary) and new records are appended
  with those four fields null. Previous export kept as
  attacks-export-Gart-website-2026-09-09-before.json; raw feed saved as
  attacks-live-api-2026-09-09.json. Export now has 389 records.
- 25 new dossiers built; 93 linked URL slots fetched (77 fetched, 8 blocked,
  4 thin, 4 video/social); 76 blocks auto-trimmed. Wayback retry for the 12
  failed slots hit archive.org rate limits (429) and is left for a later run.
- Sheet row 252 (the Phuket robbery listed in cases-missing-from-database.json)
  approved as the source for new case 476, which is that event.
- Stale dossier for renamed case 188 removed.
- Detention flag check on the new cases: 479 proposed 0->1 (forced entry plus
  assault), 468 null->0 proposed, 483 and 486 confirmed 0; review file and
  corrections JSON updated. Worklist and README carry the procedure.

verify_all.py PASSES (2776 checks, 389 case files).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-09 18:17:12 +02:00
StellarCrowandClaude Fable 5.1 93cf7ae53a docs(detention): case 67 confirmed foiled, stays 0; description correction on worklist
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-07 16:41:18 +02:00
StellarCrowandClaude Fable 5.1 5c721e4038 docs(detention): record owner rulings on 184 and 299 (stay 0)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-07 13:42:12 +02:00
StellarCrowandClaude Fable 5.1 e36ceae8d6 docs(definitions): install the paper's Appendix B glossary as the canonical definitions
merge-output/DEFINITIONS.md holds the research paper's (v5 FINAL, February 2026)
Appendix B verbatim: eleven scenario definitions and six violence categories,
plus a mapping from DB labels to glossary names and the owner's clarifications
on the Detention flag from 2026-09-06/07. The README's own paraphrase is
replaced by a pointer so there is one source of truth.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-07 06:04:36 +02:00
StellarCrowandClaude Fable 5.1 de707f8751 docs(detention): forced-entry rule; close the ten open flag items
Owner's decisions of 2026-09-07 applied:
- Rule sharpened in README "Definitions": forced entry into the victim's
  space (home, office, hotel room, car) followed by beating or holding them
  there to take what is wanted is detention, however short.
- 354, 355 (null) and 124 set to 0. 242, 273, 413, 422, 423, 215, 298 stay 1;
  the set-to-0 proposals for them are withdrawn under the rule.
- corrections-approved.md, detention-flag-review.md and the corrections JSON
  carry the decisions; the worklist item says no open items remain.
- Summaries 242 and 273 relabelled "Detention" with the forced-entry wording.

verify_all.py PASSES.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-07 06:00:54 +02:00
StellarCrowandClaude Fable 5.1 439e83a4cb feat(sources): Wayback pass, automatic trimming, detention rule, ten cases re-sourced
Second and third rounds on the dossier pipeline, reviewed locally before commit.

Retrieval
- wayback_pass.py: retries every BLOCKED/DEAD/THIN/ERROR linked source through
  the Wayback Machine; blocks carry the snapshot date, archive URL and the live
  verdict. 143 slots retried: 93 fetched, 37 no snapshot, 12 script shells, 1 PDF.
- fetch_sources.py: fetch_wayback() helper.
- add_extra_sources.py: files staged extra_* finds (web search, not on the DB
  record) into the store with a provenance note.
- Cases re-sourced by web search: 248 (Oslo: Document.no, Avisa Oslo, NRK; the
  linked Le Parisien piece is case 242 and is marked OFF-CASE), 258 Kharkiv,
  260 Singapore, 358 Bangkok, 365 Las Vegas, 388 Verneuil-sur-Seine, 390 Zoersel,
  430 Homestead. 341 of 364 cases now hold a fetched article; 11, 66 and 399
  have no public text source (podcast, police video, direct victim report).

Trimming
- auto_trim.py: case-anchored furniture cut for UNTRIMMED blocks. Finds the body
  run that mentions the case, merges across subheadings and short furniture
  gaps, drops teasers, share bars, date/URL/caption lines and subscription
  pitches; refuses pages with no record term or almost no body. 302 blocks
  trimmed; 16 left on auto-trim-review.md (7 OFF-CASE suspects). Every cut is
  labelled AUTO-TRIMMED, UNREVIEWED in the dossier; the full extract stays in
  staging/. Decisions with anchors in trim-decisions-auto.json.
- build_cases.py / verify_all.py: banners for auto-trimmed and archived blocks,
  staged-file check extended to Wayback blocks, flag legend under the record.

Detention rule
- README "Definitions": the DB field `kidnappings` is the DETENTION violence
  type (victim, guard, staff or relative held to force submission or execute
  the theft), distinct from the Kidnapping scenario (taken away and held).
- detention-flag-review.md / detention-flag-corrections.json: 40 records
  reviewed with evidence; 30 set-to-1 proposals accepted by the owner on
  2026-09-06 (listed in corrections-approved.md), 10 still open.
- apply_detention_wording.py: "Violence Used" in the 70 reviewed summaries now
  names detention explicitly (62 of 70 labelled), supported by the summary's
  own text; idempotent; supersedes the batch scripts' wording.

Worklist and docs
- sofia-worklist.md: 248 decisions, detention rule item replacing the old
  "no abduction" item, fresh-search section for the textless cases.
- README-START-HERE.md: progress notes, run commands, next steps.
- Bug fixed in passing: Wayback blocks stored in-memory text with carriage
  returns; now stored as read back from disk.

verify_all.py PASSES (2321 checks).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-06 19:36:45 +02:00
StellarCrowandClaude Fable 5.1 499fd61f79 feat(sources): fetch linked URLs for the 252 spreadsheet-only cases
Until now only cases 357+ (plus a few older ones) had their linked URLs
fetched; the other 252 dossiers held just the text pasted into the
spreadsheet, often only the lede. This pass fetches every linked URL for
those cases so the full text is on file before link rot takes it.

- stage_all.py: new --include-sheet-cases flag that lifts the skip on
  cases holding spreadsheet articles. Resumable as before.
- staging/: raw extracts and index.json for the 252 cases.
- sourced-articles.json, case-files-manifest.json: 188 new UNTRIMMED
  blocks. 179 cases gained at least one fetched article; 73 got none
  (49 BLOCKED, 18 THIN, 9 DEAD, 4 video/social, 3 ERROR).
- sources/*.txt: rebuilt with build_cases.py --all. Sheet pastes are
  kept; fetched blocks sit beside them under the NOT YET TRIMMED banner.
  Files previously committed with CRLF are now LF.
- wayback-availability.json: Wayback snapshots exist for 73 of the 98
  failed URL slots; input for an archive-fetch pass.
- README-START-HERE.md: progress note and next step.

verify_all.py PASSES (1836 checks).

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GZZENdTLzNGsbNy4DyF1yt
2026-09-05 15:02:39 +02:00