Model Router

Model Router

Use this skill when a task needs an explicit model, reasoning level, or subagent-role recommendation. It is a planning aid. It does not change the current session model and it does not manage provider credentials.

Routing Questions

Before recommending a route, classify:

• task type: research, proof, implementation, review, extraction, or synthesis
• risk: low, ordinary, correctness-critical, or high-stakes
• context size: small, repo-scale, paper-scale, or multi-source
• latency sensitivity
• verification available: tests, computations, source checks, or human review
• whether work can be delegated safely

General Guidance

• Use stronger reasoning for proof, algorithms, security, correctness audits, literature synthesis, and ambiguous multi-source research.
• Use faster or smaller workers for bounded extraction, file inventory, and low-risk parallel exploration.
• Use implementation-focused workers for code changes and tests.
• Use read-only explorer roles for scoped codebase questions.
• Do not spawn agents unless the user explicitly asks for multi-agent or delegated work.

Recommendation Format

Return:

• recommended role or agent type
• recommended model tier or reasoning level using the current system's available options
• why that route fits the task
• verification that should gate the result
• fallback if the preferred model or role is unavailable

Guardrails

• Treat the current tool definitions and agent runtime as the source of truth.
• Avoid provider-specific assumptions unless the user asks for a specific provider.
• Do not recommend changing authentication, provider config, hooks, or MCP servers from this skill.

Recommended templates

When this skill is involved, consider these workflow templates (install via the workflow-templates artifact profile, or --with-deps to pull backing skills):

• autonomous-research-loop-runbook -- Bounded autonomous research-loop runbook with four stop conditions, single-path solving, mandatory cross-agent verification, fresh-agent backtracking, and five-lane broker-routed heavy-compute offload with per-lane safety gates.
• cross-agent-adversarial-review -- Producer-never-confirmer adversarial review of a paper, proof, or code artifact across agent families with a fresh-agent confirmation gate.
• engineering-delivery-loop-runbook -- Bounded build-and-deliver loop runbook: single-path implementation with seen-to-fail proof, cross-agent diff verification, behavior-preserving cleanup, and five-lane broker-routed heavy-compute offload with per-lane safety gates.
• reversible-decision-memo -- Evidence-grounded decision record with named alternatives, source-cited rationale, reversibility class and trip-wires, and a fresh-context adversarial confirmation before the decision stands.