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

merge-output/add_extra_sources.py

@@ -0,0 +1,69 @@
+"""
+File staged extra_* sources (found by web search, NOT on the DB record) into the article
+store as UNTRIMMED blocks, so build_cases.py renders them and auto_trim.py can cut them.
+
+Reads staging/<case>/index.json; every extra_N entry that is not yet in the store becomes
+a block. FETCHED entries carry the staged text, hashed; anything else (THIN, BLOCKED, DEAD,
+NOT AN ARTICLE) is filed as a reference with its verdict. Each block says in its note that
+the URL is not on the record and when it was found, so provenance stays visible.
+
+Usage: python add_extra_sources.py 258 260 ...   then auto_trim.py <cases>, build, verify.
+"""
+
+import hashlib
+import json
+import sys
+from datetime import date
+from pathlib import Path
+
+BASE = Path(__file__).resolve().parent
+STAGING = BASE / "staging"
+STORE = BASE / "sourced-articles.json"
+
+NAMES = {"document.no": "Document.no", "alt.no": "Avisa Oslo (alt.no)", "nrk.no": "NRK", "ao.no": "Avisa Oslo",
+         "mothership.sg": "Mothership", "khaosodenglish.com": "Khaosod English", "8newsnow.com": "8 News Now (KLAS)",
+         "thedefiant.io": "The Defiant", "actu17.fr": "Actu17", "lagazette-yvelines.fr": "La Gazette en Yvelines",
+         "mantes-actu.net": "Mantes Actu", "vrt.be": "VRT NWS", "nnieuws.be": "NNieuws", "indegazette.be": "In de Gazette",
+         "justice.gov": "US Department of Justice", "decrypt.co": "Decrypt", "fortune.com": "Fortune",
+         "ukrinform.ua": "Ukrinform", "hromadske.radio": "Hromadske Radio", "sud.ua": "Sud.ua", "atn.ua": "ATN"}
+
+
+def main():
+    cases = [int(a) for a in sys.argv[1:]]
+    store = json.loads(STORE.read_text(encoding="utf-8"))
+    today = date.today().isoformat()
+    added = 0
+    for cid in cases:
+        idx_path = STAGING / f"{cid:03d}" / "index.json"
+        if not idx_path.exists():
+            print(f"case {cid}: no staging index"); continue
+        blocks = store.setdefault(str(cid), [])
+        have = {b.get("field") for b in blocks}
+        for e in json.loads(idx_path.read_text(encoding="utf-8")):
+            if not e["field"].startswith("extra_") or e["field"] in have:
+                continue
+            src = NAMES.get(e["host"], e["host"])
+            note = f"NOT on the DB record. Found {today} by web search because the linked source(s) gave no usable text."
+            if e["verdict"] != "FETCHED":
+                blocks.append({"field": e["field"], "verdict": e["verdict"], "url": e["url"], "source": src,
+                               "note": note + (" " + e["note"] if e.get("note") else "")})
+                added += 1
+                continue
+            path = next(STAGING.joinpath(f"{cid:03d}").glob(f"{e['n']:02d}_*.txt"))
+            text = path.read_text(encoding="utf-8")
+            blocks.append({"field": e["field"], "verdict": "UNTRIMMED", "url": e["url"], "source": src,
+                           "language": "unknown", "text": text, "chars": len(text),
+                           "sha256": hashlib.sha256(text.encode("utf-8")).hexdigest(),
+                           "staged_file": path.name,
+                           "paragraphs_kept": f"ALL {len(text.split(chr(10) * 2))} paragraphs, untrimmed",
+                           "method": "raw HTTP retrieval, deterministic extraction",
+                           "fetched_at": e.get("fetched_at"), "http": e.get("http"), "note": note,
+                           "cut_note": "NOT TRIMMED. Full page extract; trim on read-through."})
+            added += 1
+            print(f"case {cid}: + {e['field']} {src} {len(text):,} chars")
+    STORE.write_text(json.dumps(store, ensure_ascii=False, indent=2), encoding="utf-8")
+    print(f"added {added} block(s)")
+
+
+if __name__ == "__main__":
+    main()