| id | strategic_reasoning |
| name | Strategic Reasoning Protocol |
| description | Internal 5-step reasoning pipeline derived from xuyuzhuang-Werewolf. Agent performs Situation-Suspicion-Reflection-Strategy-Response analysis implicitly before generating any output.
|
| tags | ["reasoning","core","strategy"] |
| targets | ["character_agent","vote_prompt","night_action"] |
| priority | 10 |
| dependencies | [] |
| conflicts | [] |
| behavioral_rules | ["Always base accusations on observable evidence (voting patterns, contradictions, behavior changes), never on meta-game reasoning","When making strategic decisions, weigh both short-term survival and long-term faction advantage"] |