Send a draft to GPT (OpenAI) for an independent critique. Returns ranked issues, each with a severity (critical, major, minor, nit) and a suggested fix. Call this when the user asks for a second opinion, review or critique, or before you finalise high-stakes work (code, a plan, an argument, an important message) where an outside reviewer could catch mistakes you missed. Do not call it for short or casual replies. Pass the complete draft text; GPT sees nothing else from the conversation, so include needed context in the draft or rubric. Afterwards, tell the user which issues you accept and fix, and which you reject and why.

Start a structured debate where GPT argues against your position. Returns a debate_id and GPT's opening case against you. Call this when the user asks you to debate or stress-test a position, or when a question is contested and adversarial pressure would expose weak points better than a one-off critique. State your own position on the question first (in your reply and in claude_position). Then, for each round, write your rebuttal in the chat and call debate_turn. Finish with debate_verdict. rounds is the number of rebuttal rounds after the openings (1 to 5; default 3).

Send your rebuttal in an open debate and get GPT's counter-rebuttal plus the round status. Call this after debate_open or a previous debate_turn, once you have written your rebuttal to GPT's last argument in the chat. Pass the same rebuttal text you showed the user. Respond to GPT's specific disagreements; concede points you can't defend. When no rounds remain, call debate_verdict instead.

End a debate and get a ruling from a fresh, blind GPT moderator that sees only the transcript, with the two sides relabelled A/B in random order and model names removed. Returns what is settled (and which side each point favours), what is still unresolved and what would settle it, and which side made the stronger case. Call this when all rounds are done, or earlier if the debate has stopped producing new arguments. Calling it again returns the same verdict. Report the result to the user even-handedly, including points that went against you.

Get GPT's independent answer to a question, then a blind comparison that merges it with your answer: agreements, conflicts (and which side is better supported), points only one answer has, and a merged answer. GPT answers first without seeing your answer; a second, blind call compares the two as Answer A/B in random order. Call this for questions where an independent second answer adds value: factual or technical questions with real stakes, estimates, recommendations, or when the user asks what another model thinks. Write your complete answer first and pass it as claude_answer. This makes two GPT calls, so it is slower than the other tools.