Add pre-genesis duplicate-title gate against proposals/ history

Create a new gate file duplicate_title_gate.js exporting check({title, target_files}) that loads brain/proposals/digest-*.json (cached at cycle start), computes normalized-title trigram Jaccard vs every prior proposal's title, AND set-intersection of target_files. Reject if Jaccard >=0.7 OR target_files set-intersection >=2 with any prior status in {APPROVE, DEFER}. Wire into genesis.js between candidate generation and filter_score: any rejected proposal gets stamped {kill_reason:'duplicate_with_prior'} and skips filter_score+argument (preserves cost). Emit duplicate_count to cycle telemetry.

On the next 7 cycles, duplicate_with_prior kills should account for 10-30% of pre-filter rejects (currently invisible because dupes pass filter_score and burn argument tokens). Average tokens-per-graduated-proposal should drop by >=8% over a 7-cycle window.

If duplicate_with_prior kill rate is <5% across 7 cycles, the gate is unnecessary (prior dedup elsewhere already covers it); if >50%, the threshold is too aggressive and over-rejecting legitimate iterations.

Rejected at the council verdict. The two-judge council did not find the case strong enough to advance to Commander review.