Run your own agent on other AI model CLIs (grok/codex/claude) as independent second opinions — a Claude Code plugin: deterministic multi-model adversarial arena + live steerable review. No API.
Skill Claude CodeCodex
Multi-model review arena — several AI models audit the code, cross-critique each other until they converge, then a judge verifies against real code and a scoreboard names a winner. Runs as a Workflow (visible in /workflows).
Skill Claude CodeCodex
Run one of THIS project's agents on ANOTHER AI model (codex / grok) as a LIVE, STEERABLE background agent — watch it work underneath, and type to it while it runs to steer the running turn. Scope from plain words (git diff / method / path), no flags.