trap-detectorlisted

# Trap Detector Skill Version: 0.1 ## Purpose This skill detects whether the current movement is entering a trap. A trap is not an emotional state and not a moral failure. A trap exists when movement appears to continue, but the structural quality of the path is degrading in a way that increases distortion, false certainty, local optimization, or irrecoverability. ## When to use Use this skill when one or more of the following appear: - the user is looping without new gain - the conversation is accelerating but becoming less interpretable - a decision is forming too early - abstraction is increasing while actionable clarity is decreasing - output is being requested before structural grounding exists - the user appears stuck between multiple unresolved threads - execution pressure is outrunning coherence - a proxy goal may be replacing the real goal Do not use this skill for ordinary uncertainty or early-stage open exploration. ## Input expectations Possible inputs: - current user message - recent thread history - prior structural state - current phase / move_type if available - execution pressure indicators - translation context - decision context ## Output contract Return: - `trap_flag` - `trap_type` - `trap_confidence` - `distortion_vector` - `evidence` - `recommended_response` Example: ```json { "trap_flag": true, "trap_type": "premature_closure", "trap_confidence": "medium", "distortion_vector": "decision speed exceeds structural grounding", "ev