Skip to content
All Skills

Native-First Escalation Routing

Route ordinary init, review, and security requests to Claude-native capabilities first; escalate to Octopus only when multi-LLM diversity adds value

Agent Orchestration|v1|Updated 5/18/2026|GitHub source
MCP get_skill({ skillId: "native-first-escalation-routing-2cbe0fd7" })

Use this skill with your agent

Create a free account and connect via MCP

Get Started Free
> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex.
> Cross-reference commands use installed skill names in Codex rather than `/octo:*` slash commands.
> Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it.
> For host tool equivalents, see `skills/blocks/codex-host-adapter.md`.


# 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, Gemini, 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 Gemini 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.
#github#broad-capability#external#license-mit#nyldn-claude-octopus#agent-teams#workflow#skill#native#escalation#routinggithubanthropic-apifilesystem-access

Related Skills

More skills in Agent Orchestration

1. Product type search — what design patterns fit this product?

Design UI/UX systems with style guides, palettes, typography, and component specs for new interfaces

#github#broad-capabilityMIT

Acp Router

Route plain-language requests for Claude Code, Cursor, Copilot, OpenClaw ACP, OpenCode, Gemini CLI, Qwen, Kiro, Kimi, iFlow, Factory Droid, Kilocode, or explicit ACP harness work into either OpenClaw ACP runtime sessions or direct acpx-driven sessions ("telephone game" flow). For coding-agent thread requests, read this skill first, then use only `sessions_spawn` for thread creation. Codex chat binding defaults to the native Codex app-server plugin unless ACP is explicit or background spawn needs ACP.

#broad-capability#browserMIT

Add New Safe Output Type

Adding a New Safe Output Type to GitHub Agentic Workflows

#github#broad-capabilityMIT

Agent Governance

Patterns and techniques for adding governance, safety, and trust controls to AI agent systems. Use this skill when: - Building AI agents that call external tools (APIs, databases, file systems) - Implementing policy-based access controls for agent tool usage - Adding semantic intent classification to detect dangerous prompts - Creating trust scoring systems for multi-agent workflows - Building audit trails for agent actions and decisions - Enforcing rate limits, content filters, or tool restrictions on agents - Working with any agent framework (PydanticAI, CrewAI, OpenAI Agents, LangChain, AutoGen)

#github-copilot#agentMIT

Agentic Development

Build AI agents with Pydantic AI (Python) and Claude SDK (Node.js)

#claude-bootstrap#bootstrapMIT

Agentic Workflows

Route gh-aw workflow design/create/debug/upgrade requests to the right prompts.

#broad-capability#agentic-workflowsMIT