_agent-evaluation-project

To evaluate agent scope, you must verify that each agent definition has a distinct role, clear triggers, and explicit out-of-scope boundaries to prevent misrouting and scope creep within your system architecture.

Validating agent routing involves analyzing project-context use cases to ensure the description guides correct delegation, verifies frontmatter requirements like name and tools, and confirms boundaries are clear.

Scope creep and misrouting occur when agent definitions lack a clear purpose or have overlapping responsibilities, causing misalignment between the stated triggers and the actual project-context use cases.

Yes, you can use project-context use cases to verify frontmatter requirements such as name, description, and tools, ensuring proper delegation and boundary clarity align with system architecture constraints.

The best way to identify agent purpose misalignments is to review the agent definition against project requirements, checking if the agent claims overlapping roles like data collection and decision-making simultaneously.

You should not use a single agent for multiple tasks when the responsibilities span different architectural layers, as combining data collection and decision-making violates boundary clarity and causes improper delegation.