Three-tier constraint model, mode-aware eval, boundary examples, playtest tooling

Eval harness:
- Mode-aware scoring: sudo=strict (exact match), pray/god=soft (category match,
  in-character, appropriate intensity)
- New metrics: cmd_category_match, appropriate_intensity, scoring_mode breakdown
- Eval defaults to steel141 (192.168.0.141) — prod GPU reserved for serving

Dataset (213 examples):
- Added 31 boundary/adversarial examples (safety edges, abstention, near-boundary)
- Updated pray example reasoning: character-driven logic, not prescriptive outputs
- Tagged pray examples with scoring_mode=soft

Playtest tooling:
- whitelist.sh: add/remove/list across all 3 servers
- FRIENDS_INVITE.md + Discord version: playtester recruitment docs
- Server addresses and implementation details for both training servers

PLAN.md:
- Three-tier constraint model documented (sudo/pray/god_system)
- Success criteria split by scoring mode
- All session decisions logged

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

eval/live_bakeoff.py

@@ -522,7 +522,7 @@ def main():
     parser.add_argument("--models", nargs=2, default=["gemma3n:e4b", "qwen3:8b"],
                         metavar=("MODEL_A", "MODEL_B"),
                         help="Two models to compare (default: gemma3n:e4b qwen3:8b)")
-    parser.add_argument("--ollama-url", default="http://192.168.0.179:11434",
+    parser.add_argument("--ollama-url", default="http://192.168.0.141:11434",
                         help="Ollama API URL")
     parser.add_argument("--rcon-host", default="192.168.0.244",
                         help="RCON host (default: 192.168.0.244)")