What makes this different from a standard AI code review?

Unlike standard single-pass reviews, this skill uses multiple specialized agents in parallel—focusing on specific domains like security, performance, and architecture—to provide a much deeper level of scrutiny.

Is there a limit to how many times it will try to fix an error?

The skill is configured with a maximum of 5 iterations and includes stall detection to prevent infinite loops if the AI is unable to resolve a specific finding.

Can it automatically fix the issues it discovers?

Yes, the skill includes an iteration loop that identifies issues, applies suggested fixes via the Edit tool, and re-reviews the code until the issues are resolved or the iteration limit is reached.

Is it safe to use for security-sensitive code?

While it automates fixes, the skill includes a safety protocol where high-severity findings are flagged for orchestration review, preventing the blind application of LLM-suggested fixes to sensitive areas.

How does it decide which specialist agents to use?

It uses built-in signal detection to analyze the files being reviewed. For example, if it detects SQL or schema files, it automatically activates the Database Specialist pass.

Orchestrate Review - Claude Code Skill for Multi-Pass Audits

概要

The Orchestrate Review skill transforms standard AI code reviews into a robust, multi-pass auditing system. By spawning parallel specialized agents for code quality, security, performance, and test coverage, it identifies deep-seated issues that single-pass reviews often miss. It features intelligent signal detection to automatically pull in domain-specific experts—such as database, API, or frontend specialists—based on the codebase context. Most importantly, it includes an automated iteration loop that applies fixes and validates them until the code meets high-quality standards or reaches an iteration threshold.

主な機能

Automated fix-and-verify iteration loop with stall detection
Multi-pass parallel review architecture using specialized domain agents
Structured JSON findings reporting with severity and confidence scores
Context-aware signal detection for automated specialist selection
Smart aggregation and deduplication of cross-pass review findings
311 GitHub stars

ユースケース

Standardizing code quality and pattern consistency across large-scale repositories
Automating the final review and 'deslop' phase of a complex development cycle
Performing deep security and performance audits on critical feature pull requests

概要

主な機能

Automated fix-and-verify iteration loop with stall detection
Multi-pass parallel review architecture using specialized domain agents
Structured JSON findings reporting with severity and confidence scores
Context-aware signal detection for automated specialist selection
Smart aggregation and deduplication of cross-pass review findings
311 GitHub stars

ユースケース

Standardizing code quality and pattern consistency across large-scale repositories
Automating the final review and 'deslop' phase of a complex development cycle
Performing deep security and performance audits on critical feature pull requests

Multi-Pass Code Review Orchestrator

Multi-Pass Code Review Orchestrator

概要

主な機能

ユースケース

概要

主な機能

ユースケース