skill-native-escalation-routing

name	skill-native-escalation-routing
disable-model-invocation	true
description	Use when choosing native or multi-LLM handling for init, review, or security requests

Native-First Escalation Routing

Use this skill when the user asks for repository initialization, code review, or security review and it is not yet clear whether Claude-native behavior is sufficient or whether Octopus escalation is warranted.

Core Policy

Claude-native first:

• /init
• /review
• /security-review

Octopus for escalation:

• /octo:review
• /octo:security
• /octo:debate
• /octo:multi

Route to Claude-Native First

Prefer Claude-native behavior when all of the following are true:

• the request maps directly to init, review, or security review
• the user did not ask for multiple model opinions
• the user did not ask for Codex, Antigravity, or another provider explicitly
• the task does not require adversarial debate, consensus scoring, or external-provider specialization

Examples:

• "initialize this repo"
• "review my staged changes"
• "security review this auth module"

Escalate to Octopus

Escalate when the user asks for or clearly benefits from:

• multiple model opinions
• adversarial review
• debate between providers
• provider-specific analysis
• autonomous codegen verification
• elevated rigor for complex or high-risk work

Examples:

• "get multiple model opinions on this PR"
• "have Codex and Antigravity review this architecture"
• "do an adversarial security review"
• "verify this AI-generated implementation before merge"

Execution Guidance

• If Claude-native is sufficient, do not wrap the task in Octopus unnecessarily.
• If the user explicitly invokes /octo:review or /octo:security, treat that as an escalation request.
• If the request is ambiguous, prefer Claude-native first and mention Octopus as the escalation path.

Suggested User Framing

Use wording like:

Claude-native first, Octopus for escalation. Use Claude-native /review or /security-review for ordinary requests. Use Octopus when you want multiple model opinions, adversarial review, or stricter multi-LLM workflows.