skill-native-escalation-routing

Code review tasks route to Claude-native capabilities by default using the /review command. This ensures ordinary repository initialization and security review tasks execute efficiently without unnecessary multi-LLM overhead.

Escalate an AI workflow to multi-LLM adversarial review when you need multiple model opinions, higher rigor, or strict multi-LLM diversity that adds value beyond standard Claude-native code review processing.

Multi-LLM routing governance for repository initialization is managed by defaulting to Claude-native commands like /init, escalating to Octopus only when strict multi-LLM workflows are specifically beneficial for the task.

Yes, you can use Octopus escalation commands for security review and multi-model debates. The /octo:security and /octo:debate commands specifically route tasks to controlled multi-LLM workflows.

Claude-native routing supports repository initialization without external dependencies. The /init command executes locally, providing a lightweight routing governance approach before escalating to multi-LLM workflows.

Code review needs escalation to Octopus instead of standard routing when standard Claude-native processing lacks the multi-LLM diversity or adversarial review rigor required for complex security analysis tasks.