feat: add value-violations adversary suite (#617)

training/README.md

@@ -20,6 +20,7 @@ make train-local    # LoRA on Apple Silicon (MLX) — free, ~30 min on M3 Max
 make train-cloud    # QLoRA on cloud GPU (Axolotl) — ~$1/run on A100
 make eval           # Standard benchmarks via lm-eval-harness against Ollama
 make vibes          # Hand-picked prompts → human review (the sacred test)
+make adversary-value-violations  # 200 jailbreak prompts targeting value violations
 make ingest         # Pull heartbeat trajectories into training data
 make curated        # Regenerate curated exemplar dataset
 make convert        # Convert merged data to MLX train/valid format
@@ -46,10 +47,12 @@ training/
 ├── eval-tasks.yaml          ← Benchmark config
 ├── build_curated.py         ← Exemplar data authoring (the soul conversations)
 ├── ingest_trajectories.py   ← Quality filter for heartbeat cycle data
+├── run_adversary_eval.py    ← Adversarial prompt runner + transcript capture
 └── data/
     ├── curated_dataset.jsonl       ← 26 gold-standard conversations (proprietary)
     ├── preference_pairs.jsonl      ← DPO preference pairs (proprietary)
     ├── prompts_vibes.yaml          ← Custom eval prompts
+    ├── prompts_adversary_value_violations.yaml ← 200 jailbreak prompts for value-violation red-team eval
     ├── prompts_nexus_vibes.yaml    ← Nexus-specific eval prompts
     └── mlx_curated/                ← MLX-format train/valid splits
 ```