What happens if a Medium-severity issue is identified?

Medium issues require a mandatory fix and a full re-review of all three agents before the gate can be passed. They cannot be deferred with comments.

Can I skip the review for trivial code changes?

No. Because security issues can reside in even a single line of code, this skill enforces a 'no exceptions' policy for all implementation changes.

Can I run the reviewers sequentially to save time?

No, sequential execution is strictly forbidden. The skill requires parallel dispatch in a single message to ensure a holistic feedback loop and minimize total review time.

Why does this skill require three separate reviewers?

It uses specialized agents for code, business logic, and security to provide deep, domain-specific insights that a single general-purpose agent might overlook.

What does the NEEDS_DISCUSSION verdict mean?

This indicates a reviewer is uncertain and requires a human decision. The gate remains blocked until the user provides guidance and the reviewer is re-run.

Dev Review (Gate 4)

Name: Dev Review (Gate 4)
Author: LerianStudio

byLerianStudio

•

セキュリティとテスト

Orchestrates mandatory parallel code reviews using three specialized AI agents to enforce security, logic, and quality standards.

概要

The Dev Review skill automates the critical Gate 4 phase of the development cycle by dispatching specialized reviewers for code, business logic, and security in a single parallel message. It eliminates the 'rubber-stamp' effect by requiring three distinct perspectives to catch vulnerabilities and logic flaws that general-purpose agents often miss. Designed with high pressure resistance, this skill enforces strict adherence to engineering best practices, ensuring that Medium, High, and Critical issues are resolved before any code can proceed to validation or deployment.

主な機能

17 GitHub stars
Standardized issue severity mapping and resolution protocols
Automated verdict aggregation (PASS, FAIL, NEEDS_DISCUSSION)
Parallel 3-agent review (Code, Business, and Security specialists)
Comprehensive documentation of findings with file and line references
Mandatory gate enforcement for Medium-severity issues and above

ユースケース

Automating security and business logic verification for production code
Standardizing multi-perspective code audits in AI-driven dev cycles
Enforcing quality gates that prevent technical debt and security regressions

概要

主な機能

17 GitHub stars
Standardized issue severity mapping and resolution protocols
Automated verdict aggregation (PASS, FAIL, NEEDS_DISCUSSION)
Parallel 3-agent review (Code, Business, and Security specialists)
Comprehensive documentation of findings with file and line references
Mandatory gate enforcement for Medium-severity issues and above

ユースケース

Automating security and business logic verification for production code
Standardizing multi-perspective code audits in AI-driven dev cycles
Enforcing quality gates that prevent technical debt and security regressions