How often should I run the verification loop?

It is recommended to run it after completing a specific function or component, every 15 minutes during long sessions, or as the final step before creating a pull request.

What languages does the Verification Loop support?

It provides built-in configurations for TypeScript/JavaScript (using npm/pnpm/npx) and Python (using pyright/ruff), but its logic can be adapted for other environments.

Does this skill catch security issues?

Yes, it includes a dedicated security scanning phase that checks for hardcoded API keys (like Anthropic 'sk-' keys) and accidental console logs in production source code.

How does this differ from standard CI/CD pipelines?

While similar to CI, this skill provides immediate, local feedback within the Claude Code session, allowing the AI to identify and fix issues before the code is even committed to the repository.

What is the target test coverage for this skill?

The skill establishes a target of 80% minimum test coverage to ensure that all changes are backed by reliable automated tests.

Verification Loop

Name: Verification Loop
Author: oabdelmaksoud

byoabdelmaksoud

0•

安全与测试

Automates a comprehensive multi-stage validation process to ensure code quality, security, and stability before finalizing development tasks.

Verification Loop is a robust quality assurance framework designed for Claude Code sessions, providing a standardized workflow for validating software changes. It systematically guides Claude through six critical stages: build verification, type checking, linting, test suite execution with coverage targets, security scanning for secrets, and detailed diff reviews. By enforcing these quality gates, the skill ensures that AI-generated code is production-ready, maintains high test coverage (80%+ target), and remains free of common security vulnerabilities or unintended side effects before a pull request is created.

主要功能

01Cross-language support for TypeScript/JavaScript and Python ecosystems

020 GitHub stars

03Multi-stage verification including Build, Type, Lint, Test, and Security phases

04Security auditing for hardcoded secrets, API keys, and debug logs

05Automated test coverage reporting with enforced 80% minimum thresholds

06Standardized verification report generation for clear PR-readiness status

使用场景

01Validating features or bug fixes immediately before submitting a pull request

02Ensuring large-scale refactorings haven't introduced regressions or type errors

03Maintaining consistent code quality during long-running AI development sessions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add oabdelmaksoud/superclaw verification-loop

For use in Claude.ai and ChatGPT

主要功能

01Cross-language support for TypeScript/JavaScript and Python ecosystems

020 GitHub stars

03Multi-stage verification including Build, Type, Lint, Test, and Security phases

04Security auditing for hardcoded secrets, API keys, and debug logs

05Automated test coverage reporting with enforced 80% minimum thresholds

06Standardized verification report generation for clear PR-readiness status

使用场景

01Validating features or bug fixes immediately before submitting a pull request

02Ensuring large-scale refactorings haven't introduced regressions or type errors

03Maintaining consistent code quality during long-running AI development sessions