buildvsbuy-ai.vercel.app

solve_winning_conditions

Sensitivity analysis: what must change for target_option_id to beat the current math leader? NOT for vibe-coding / prototype-speed / "does that change who you'd recommend" — those are set_priority_weight(time_to_prototype, 9 or 10) which auto-reranks. On the auth preset, this tool is REFUSED until a real set_priority_weight has run this session. After priorities are set: computes flip_levers (0-10, step 0.1), score_gap, and suggested_context_shifts. Never assert flip-lever claims without calling this tool.

Input schema

{
  "type": "object",
  "properties": {
    "target_option_id": {
      "type": "string",
      "description": "Option id that should win."
    }
  },
  "required": [
    "target_option_id"
  ],
  "additionalProperties": false
}