What problem does it solve? It provides an independent, clean-context review of a finished or modified agent definition, checking both convention compliance and system prompt quality before the agent goes live, without ever editing the agent itself. ## Core Features & Use Cases - Convention compliance check: Validates naming rules, the mandatory frontmatter five-tuple, the model × effort × maxTurns matrix, and justification comments for omitted fields. - System prompt quality audit: Detects blocking issues such as instructions to ask the user, tools mismatched with the mandate, references to project documents or .claude/rules, and unanchored paths. - Structured verdict output: Returns an approved/rework verdict with findings classified as blocking, recommendation, or note, including verification of unresolved findings from previous review rounds. - Use Case: After write-agent finishes a round of edits on a new agent, it spawns this reviewer as a subagent to independently confirm the agent meets conventions before approval. ## Quick Start Ask the assistant to review the agent at a given path against the plugin conventions and return the verdict with findings.
