Does it support custom security rules?

Yes, it supports Policy-as-Code via a JSON configuration file where you can define minimum thresholds for security, coverage, and quality that can block merges if requirements are not met.

Which agents are involved in the verification process?

The process involves specialized agents for code quality, security auditing, test generation, backend architecture, frontend development, and performance engineering.

What are the 8 phases of verification?

The workflow includes context gathering, parallel agent dispatch, test execution, nuanced grading, improvement suggestions, alternative comparisons, metrics tracking, and final report compilation.

What is OrchestKit Verify?

OrchestKit Verify is a specialized skill for Claude Code that automates the verification of code changes using a multi-agent orchestration approach to ensure high-quality, secure, and performant implementations.

How does the grading system work?

The skill uses six specialized agents to evaluate different dimensions of your code, providing a 0-10 score for each and calculating a final composite grade from A (Excellent) to F (Blocked).

OrchestKit: Verify

Name: OrchestKit: Verify
Author: yonatangross

byyonatangross

•

보안 및 테스팅

Conducts multi-layered verification of code implementations using parallel specialized AI agents and nuanced grading.

OrchestKit Verify is a high-complexity validation skill for Claude Code that orchestrates a team of six specialized agents to audit security, performance, code quality, and UI/UX. It follows an automated 8-phase workflow—from context gathering to final report generation—calculating composite scores and letter grades to determine if a feature is production-ready. By utilizing parallel task execution and adaptive thinking, it provides developers with prioritized improvement suggestions and detailed evidence-based artifacts for every code change, ensuring high-standard implementations through policy-as-code thresholds.

주요 기능

01Parallel multi-agent auditing for security, quality, performance, UX, API, and testing.

02Comprehensive report generation with prioritized effort-vs-impact improvement suggestions.

03Nuanced 0-10 grading system with composite scoring and letter grades (A-F).

04Policy-as-code support for defining custom quality and security gate thresholds.

0592 GitHub stars

06Automated 8-phase verification workflow with Git diff and history integration.

사용 사례

01Performing pre-merge security and quality audits on complex feature branches.

02Ensuring React 19 frontend implementations meet accessibility and design system standards.

03Validating backend API migrations and database changes for performance regressions.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit verify

For use in Claude.ai and ChatGPT

Download Skill