Route subagents to the cheapest sufficient model across task types.
npx skills add https://github.com/ItsMattG/property-tracker --skill model-routing
Please help me install this Agent Skill. Skill: model-routing Source: https://github.com/ItsMattG/property-tracker/tree/main/.claude/skills/model-routing Command: npx skills add https://github.com/ItsMattG/property-tracker --skill model-routing
Route subagents to the cheapest sufficient model to optimize cost and resource use.
Balance cost and capability by mapping each task type to the cheapest model that can complete it.
High-intent search queries and answers about installing and using this skill.
Model routing assigns each task type to the cheapest capable model using a configurable routing table, optimizing cost and performance across autonomous workflows.
The fallback escalation path for task dispatch is haiku to sonnet to opus. This tiered routing strategy handles failures and maintains throughput by upgrading to more capable models as needed.
Cost-aware routing works by mapping task types like multi-file implementation and code reviews to specific models in a routing table, ensuring exploratory searches use cheaper models while architecture tasks use capable ones.
Yes, you can use a custom routing table for autonomous AI workflow task dispatch. The routing table is configurable, allowing you to balance cost and capability by mapping each task type to preferred models.
You should not use the cheapest model for AI task routing when handling complex architecture reviews or multi-file implementations that exceed its capability. The tiered fallback escalation to sonnet or opus handles these failures.
richardnpaulDecompose engineering tasks into verifiable units with eval-first evaluation and tiered model routing.
campbellmcgregorGuide AI-assisted software builds with workflows, model selection, and prompt strategies.
galangryandanaDispatch fresh subagents to execute independent implementation tasks with two-stage reviews.
latestaiagentsRoute LLM tasks to models based on complexity, cost, and latency.
