feat: DPO pair quality validator — gate before overnight training

Add DPOQualityValidator that catches bad training pairs before they
enter the tightening loop. Wired into DPOPairGenerator between
generate() and export() as an automatic quality gate.

New module: dpo_quality.py
- 5 single-pair quality checks:
  1. Field length minimums (prompt ≥40, chosen ≥80, rejected ≥30 chars)
  2. Chosen/rejected length ratio (chosen must be ≥1.3x longer)
  3. Chosen≈rejected similarity (Jaccard ≤0.70 — catches low-contrast)
  4. Vocabulary diversity in chosen (unique word ratio ≥0.30)
  5. Substance markers in chosen (≥2 fleet/training/action terms)
- 2 cross-pair quality checks:
  6. Near-duplicate prompts within batch (Jaccard ≤0.85)
  7. Cross-run dedup against recent JSONL history files
- Two modes: 'drop' (filter out bad pairs) or 'flag' (export with warning)
- BatchReport with per-pair diagnostics, pass rates, and warnings
- Standalone CLI: python3 dpo_quality.py <file.jsonl> [--strict] [--json]

Modified: dpo_generator.py
- Imports DPOQualityValidator with graceful degradation
- Initializes from config validation section (enabled by default)
- Validates between generate() and export() in run()
- Quality report included in pipeline result dict
- Validator failure never blocks — falls back to unvalidated export

Modified: config.yaml
- New deepdive.training.dpo.validation section with all tunable knobs:
  enabled, flagged_pair_action, similarity thresholds, length minimums,
  dedup_history_files

Integration tested — 6 test cases covering:
  ✓ Good pairs pass (3/3 accepted)
  ✓ Bad pairs caught: too-short, high-similarity, inverted signal (0/3)
  ✓ Near-duplicate prompt detection (1/2 deduped)
  ✓ Flag mode preserves pairs with warnings (3/3 flagged)
  ✓ Cross-run deduplication against history (1 dupe caught)
  ✓ Full generator→validator→export pipeline (6/6 validated)

intelligence/deepdive/config.yaml

@@ -99,6 +99,16 @@ deepdive:
         - "summarize"        # Paper summary → fleet-grounded analysis
         - "relevance"        # Relevance analysis → scored fleet context
         - "implication"      # Implications → actionable insight
+      validation:
+        enabled: true
+        flagged_pair_action: "drop"      # "drop" = remove bad pairs, "flag" = export with warning
+        min_prompt_chars: 40             # Minimum prompt length
+        min_chosen_chars: 80             # Minimum chosen response length
+        min_rejected_chars: 30           # Minimum rejected response length
+        min_chosen_rejected_ratio: 1.3   # Chosen must be ≥1.3x longer than rejected
+        max_chosen_rejected_similarity: 0.70  # Max Jaccard overlap between chosen/rejected
+        max_prompt_prompt_similarity: 0.85    # Max Jaccard overlap between prompts (dedup)
+        dedup_history_files: 5           # How many recent JSONL files to scan for cross-run dedup
 
   # Phase 0: Fleet Context Grounding
   fleet_context: