What problem does it solve? In multi-agent teams, rejected work often gets bounced back to the original author, creating defensive self-revision loops and weak independent review. This Skill defines a strict protocol so coordinators enforce reviewer rejections correctly and prevent the original author from revising their own rejected artifact. ## Core Features & Use Cases - Reviewer Rejection Protocol: Reviewers approve or reject work, and on rejection must choose to reassign to a different agent or escalate to a new agent with specific expertise. - Strict Lockout Semantics: The original author is mechanically locked out of the rejected artifact for the entire revision cycle, including as co-author or advisor, with per-artifact scope. - Deadlock Handling: When all eligible agents are locked out of an artifact, the coordinator escalates to the user instead of re-admitting a locked-out author. - Use Case: A tester agent rejects an authentication module written by another agent; the coordinator verifies the named revision agent is not the original author, spawns a different agent to produce v2, and keeps the lockout active until the revision is approved. ## Quick Start Apply the reviewer-protocol rules whenever a reviewer agent rejects an artifact so the coordinator locks out the original author and assigns the revision to a different agent.
