task-classification

What problem does it solve?

It determines how to handle an incoming request by estimating scope, risk, and verification needs so the orchestration chooses the correct agent path.

Core Features & Use Cases

• Complexity & risk assessment: Classifies tasks into Trivial, Exploratory, Implementation, Complex, or Research based on read-only status, impact size, and high-risk domains (security, auth, DB schema, billing).
• Agent routing and orchestration planning: Selects the appropriate agent sequence (Riko, Senku, Loid, Lawliet, Alphonse) or triggers full orchestration vs single-pass execution.
• Verification gate requirements: Defines which task classes must run tests, lint, integration checks, and/or security review, preventing completion without required verification.
• Parallel team eligibility guidance: Recommends parallel execution when subtasks are independent with exclusive file ownership, deferring to team-decision guidance for criteria.
• Re-classification rules: Provides a protocol to upgrade classification when scope or dependencies grow during exploration or implementation.

Quick Start

Use the task-classification skill when you want an AI to decide whether your request is exploratory, implementation, or complex and to specify which verification gates must run before changes are considered complete.