What is a manifest file in this context?

A manifest is a structured document, usually generated by a planning skill like /define, that outlines specific deliverables, acceptance criteria, and global invariants for a task.

Does this skill automatically verify the code?

Yes, the skill is designed with a mandatory constraint to call a verification process after deliverables are addressed, ensuring that all acceptance criteria are truly satisfied.

Why does the skill maintain an execution log?

The execution log acts as a persistent working memory, allowing the AI to track progress, store context, and recover state during long-running or complex development tasks.

When should I use the 'do' skill?

Use it when you have a clear plan or manifest and want the AI to systematically implement the requirements and verify the results on its own.

How does the 'do' skill handle execution failures?

It targets specific failing criteria identified during the verification phase, allowing the agent to fix isolated issues without restarting the entire implementation or touching passing code.

Manifest Task Executor

Name: Manifest Task Executor
Author: doodledood

bydoodledood

•

生产力与工作流

Executes structured development manifests by iterating through deliverables and ensuring all acceptance criteria are met through automated verification.

The 'do' skill is a high-precision execution engine designed for agentic development workflows. It takes a pre-defined manifest (typically generated by a planning tool like /define) and methodically transforms requirements into reality. By maintaining a persistent execution log and enforcing mandatory verification cycles, this skill ensures that every deliverable satisfies its specific acceptance criteria while respecting global project invariants. It is ideal for complex, multi-step coding tasks where maintaining context and minimizing rework are essential for productivity.

主要功能

018 GitHub stars

02Mandatory integration with automated verification suites

03Smart todo discipline to externalize progress and manage sub-tasks

04Automated manifest parsing and goal-oriented execution

05Targeted failure resolution focusing only on specific failing criteria

06Persistent execution logging for context recovery and state management

使用场景

01Automating the 'Build-Verify-Fix' cycle for large-scale code changes

02Executing complex feature implementations based on pre-defined architectural plans

03Ensuring strict adherence to acceptance criteria in mission-critical development tasks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add doodledood/claude-code-plugins do

For use in Claude.ai and ChatGPT

Download Skill