Routes subagent tasks to Claude models (opus, sonnet, haiku based on task type.
npx skills add https://github.com/mec07/claude_skills --skill modelrouting
Please help me install this Agent Skill. Skill: ModelRouting Source: https://github.com/mec07/claude_skills/tree/main/stow/ModelRouting Command: npx skills add https://github.com/mec07/claude_skills --skill modelrouting
Route subagents to the right model (opus/sonnet/haiku) based on task type to optimize performance and resource use.
Dispatch a task to the routing system and it will select the appropriate model (opus, sonnet, or haiku) based on the task type.
High-intent search queries and answers about installing and using this skill.
Model routing assigns subagents to opus for reasoning, sonnet for code, and haiku for mechanical tasks to optimize performance and resource use across dispatched agent tasks.
Model selection optimization maps task types to appropriate models, assigning opus for planning and reasoning, sonnet for coding and debugging, and haiku for simple lookups, defaulting to opus for ambiguous cases.
Yes, a deviation policy allows justified deviations from the default model routing when specific context warrants it, requiring a brief justification for overriding the standard task-to-model mapping.
Yes, the model routing system includes a self-improvement hook that learns from task outcomes and refines the routing table over time to improve future model selection accuracy.
Use haiku for mechanical tasks and simple lookups to conserve resources, while opus should be used for reasoning, planning, and design tasks or any ambiguous cases where heavy processing is required.
For ambiguous tasks that do not clearly match a defined category, the default routing policy favors opus to ensure high-quality reasoning and processing across unspecified task types.
InnerScopeHearingRecommend agent count, model selection, and critic-pass usage from risk signals and effort-scale baseline.
flosrnOrchestrate multi-step software tasks with planning, subagent delegation, and verification gates.
opensage-agentOrchestrate multi-agent systems with ensemble, pool, and retry patterns.
kriscendobotRoutes AI agent tasks to appropriate model tiers with fallback logic.
