How does it handle implementation failures?

If an implementation agent fails to make tests pass after a retry, the task is marked as blocked and flagged for user attention, stopping the pipeline for that specific task until resolved.

What is required before running the executing skill?

You must have a task plan (typically a task_plan.md file) located in your docs/plans/ directory, which is usually generated by a planning-specific skill.

What coding standards are enforced?

The skill enforces Red-Green-Refactor TDD cycles, conventional commit formats with explicit file paths, and checks for SOLID compliance and project-specific conventions from CLAUDE.md.

What is the 3-stage pipeline in the ShipIt skill?

The pipeline consists of Implementation (coding with TDD), Spec Review (verifying implementation matches requirements), and Quality Review (checking for SOLID principles, error handling, and conventions).

Does it support persistent state?

Yes, it uses a progress.md file to track the status of tasks and batches, allowing you to pause execution and resume later without losing your place in the plan.

ShipIt Execution Pipeline

Name: ShipIt Execution Pipeline
Author: jugrajsingh

byjugrajsingh

0•

生产力与工作流

Automates multi-stage implementation, spec review, and quality assurance pipelines for batch-processed development tasks.

ShipIt Executing is a sophisticated orchestration skill for Claude Code that transforms static development plans into an automated execution engine. It coordinates specialized subagents through a rigorous three-stage pipeline—Implementation, Spec Review, and Quality Review—ensuring that every code change satisfies requirements and adheres to high-quality standards. By managing execution state through persistent progress files and utilizing native Task coordination, it allows developers to automate complex feature builds while maintaining human-in-the-loop control for critical decision points.

主要功能

01Automated quality reporting with severity-coded findings (Minor, Major, Critical)

02Persistent state management via progress.md for session-resilient execution

03Three-stage subagent pipeline: Implement, Spec Review, and Quality Review

04Strict Red-Green-Refactor TDD enforcement for implementation agents

050 GitHub stars

06Dependency-based native Task coordination for real-time status tracking

使用场景

01Managing long-running development cycles with the ability to pause and resume execution

02Enforcing high code quality and architectural standards through automated multi-agent reviews

03Scaling complex feature development by automating the implementation of multi-task plans

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jugrajsingh/skillgarden executing

For use in Claude.ai and ChatGPT

Download Skill