What problem does it solve? Single-pass debugging often produces fragile conclusions. This Skill orchestrates a structured multi-agent debate where N independent solvers argue over R rounds and an isolated judge with a frozen rubric synthesizes one recommendation for a registered bug's diagnosis, repair strategy, or spec verdict. ## Core Features & Use Cases - Three debate modes: diagnosis (competing causal hypotheses), repair (competing fix strategies), and spec (recommendation on whether code, tests, or spec is correct). - Isolated judge with frozen rubric: the judge sees only anonymized, shuffled final proposals, preventing echo chambers and reward hacking. - Auditable state and failure handling: all rounds, convergence metrics, and the final answer are written to the bug's debate folder, with hard timeouts, quorum rules, and explicit cost estimates before running. - Optional external harnesses: CLI agents like Codex or Gemini can join as solvers or critics, but only with explicit user consent and never for restricted-visibility bugs. - Use Case: A registered bug has three plausible root causes. Run the debate in diagnosis mode with 3 solvers and 2 rounds; the judge returns a synthesized hypothesis with a discriminating probe, which you then execute in the fix workflow. ## Quick Start Ask the agent to run a multi-agent debate on bug BUG-20260715-A7K3 in diagnosis mode with the default 3 solvers and 2 rounds.
