Task Complexity Routing and Agent Selection

★ 7.7 · orchestration

task-classifier is a Claude Code skill that evaluates task complexity on a 1–10 scale and automatically determines the appropriate model, agent team composition, and execution mode before a task runs. The skill is not user-invocable — it operates as an internal orchestration component. The scale spans five tiers: trivial single-line fixes (1–2, GLM-4.7, no agents) through critical system-wide redesigns (9–10, Claude Opus, full six-agent team including coder, reviewer, tester, researcher, frontend, and security). At complexity levels 7–10, plan mode and agent involvement become mandatory, and Aristotle switches to its full five-phase workflow. The skill integrates with /orchestrator and the model routing layer to ensure every task is handled by the right resources.