Full STRIKE2022 pipeline ran clean: 63 filing extracts (3,330 claims,
56,008 total), embed-claims indexed all into Qdrant, two-sided live/test.
Engine refuses the false positive (net=+0.25, capped single bitcoin
cluster, far below the 2.0 firing bar) but the own_network-drop
reflexivity demo is unexercised (own_net=0, live==test) because
RHR/CD/Bitcoin.Review were deferred at transcription 2026-06-08.
Accepted as a conditional pass; un-defer follow-up tracked in ROADMAP.
Fold the hang lesson into the Gemini operational rule (disable thinking AND set a timeout) and refresh Current state for the in-progress Gemini extraction (~68 docs left, 52.7k claims) and the gating Strike test.
A timeout-less generate_content call hung the single-threaded extract worker for ~50 min mid-batch. Set an HTTP timeout (120s) plus 4 retries with backoff, mirroring SparkControl._post; transient 504/read-timeouts now self-heal instead of freezing the run.
Record two recurrence-prone gotchas in Key operational rules (transcript chunking must cap every chunk; Gemini backend must disable thinking) and rewrite Current state for the in-progress Gemini extraction batch and the gating Strike test.
gemini-2.5-flash thinks by default and spent ~3.8k of the 4k output budget on reasoning, hitting MAX_TOKENS with a truncated JSON body -> 0 claims parsed. Set thinking_budget=0 so the full budget goes to the answer (mirrors the local path's enable_thinking=False). On the validation chunk this went from 0 -> 11 claims.
chunk_text split only on "\n\n", but ASR transcripts have none (speaker turns are joined by a single "\n"), so whole 2-3h episodes (~250K chars) went to the extractor in one call and 400'd on context overflow. Fall through paragraph -> line -> sentence -> word -> hard char-slice so no chunk exceeds the cap regardless of punctuation; guard max_chars < 1.
Default extraction to recall-first full coverage (chunk_chars 12K, max_chunks 999) and expose both as run-extract --chunk-chars / --max-chunks.
Keysat