claude-planlisted

# Claude Plan Use this skill before substantial implementation work when a second model's decomposition could expose missed tests, hidden constraints, or a safer order of operations. Run: ```bash node "${CODEX_PLUGIN_ROOT}/scripts/claude-companion.mjs" plan "$ARGUMENTS" ``` ## Natural-Language Claude Routing <!-- routing:plan routing:codex-owns-final-plan --> - Do not ask the user to write `--quality`, `--model`, or `--effort` unless troubleshooting the plugin itself. - Default to `--quality auto` for manual Claude review, plan, rescue, and multi-review unless the command documents a stricter default. - Use `--quality strong` for deep, strict, high-risk, migration, release, or difficult diagnosis/planning requests. - Use `--quality max` only when the user explicitly asks for the strongest local Claude pass. - If the user names a concrete Claude model or effort, pass it as explicit argv tokens outside quoted `$ARGUMENTS`. - Do not substitute strong local Claude routing with `claude ultrareview`; ultrareview requires the claude-ultrareview skill and explicit cost confirmation. User-facing examples: - "Ask Claude to make an independent implementation plan." - "Use Claude for a strict migration plan review." - "Use the strongest local Claude planning pass before we edit." Internal routing procedure: - Classify the user's intent first, then invoke the narrowest Claude for Codex command that satisfies it. - If the user explicitly names Fable, Fable 5, or `claude-fable-5`, r