3 tagged agent hooks, measured the same way as everything else here.
Agent
Read-only adversarial verifier for the codebase-discovery skill. Use in the verification phase to independently check that every claim in the generated onboarding docs (under docs/) traces to real code evidence or a named stakeholder, that nothing was invented, and that the docs stay onboarding-lean. Runs as an…
Agent
Read-only code reconnaissance worker for the codebase-discovery skill. Use during deep recon to investigate a specific, scoped area of a codebase (the data model, a service, an API surface, a workflow) and return distilled, evidence-cited findings without dumping raw file contents into the main context. Ideal for…
Agent
Reviews rendered artifacts for design-contract compliance. Reports findings and recommends next action. NEVER self-renders.