What problem does it solve?
Unstructured, inconsistent code reviews lead to missed critical issues, wasted time on trivial fixes, and unclear pass/fail decisions that block pipeline progression. This skill eliminates those gaps by standardizing the review process with automatic mechanical fixes and evidence-based verdicts.
Core Features & Use Cases
- Fix-first automated corrections: Automatically resolves trivial issues like dead code, magic numbers, stale comments, and style inconsistencies without human intervention, only escalating ambiguous or high-impact changes.
- Conditional specialist review: Invokes relevant specialists (spec compliance, code quality auditor, architect, cross-runtime reviewer) based on diff scope, task mode (Express/Fast/Full), and available knowledge bases or cross-runtime CLIs.
- Structured pipeline verdicts: Generates a standardized JSON review output with GO/NO-GO status, categorized findings, and escalation flags to integrate directly with the Roster pipeline, routing work to QA, shipping, or implementation as needed.
Use case: A software team using the Roster agent pipeline can use this skill to standardize post-implementation reviews, ensure consistent quality gates across all tasks, and automatically route work based on clear, evidence-based criteria instead of ad-hoc human judgment.
Quick Start
Use the roster-review skill to conduct a structured fix-first review of the completed implementation for the current task and generate a GO/NO-GO verdict for pipeline progression.
