workflow-kit-implement

Implements a selected workflow task within managed-change workspace constraints.

Updated May 27, 2026
One-click install
npx skills add https://github.com/Kat4X/spec-kit --skill workflow-kit-implement
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: workflow-kit-implement
Source: https://github.com/Kat4X/spec-kit/tree/main/skill-drafts/workflow-kit-implement
Command: npx skills add https://github.com/Kat4X/spec-kit --skill workflow-kit-implement

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents AI-driven changes from drifting away from the intended spec by forcing implementation to be executed strictly against tasks.md, plan.md, and scope.md within a single change workspace.

Core Features & Use Cases

  • Task-first implementation: selects and executes exactly one next unchecked task (optionally by [Txxx]) from tasks.md while honoring dependencies and required confirmations.
  • Managed edit boundaries: enforces scope.md so the agent only modifies files marked as freely editable and stops before touching forbidden or high-risk surfaces.
  • Verification-driven completion: updates verification.md with concrete check results and marks tasks complete only when checks pass or clearly explain why they cannot.

Quick Start

Ask the agent to run: implement {change-name} [task-id], e.g., implement my-change T005.

Frequently Asked Questions about workflow-kit-implement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute a specific task from tasks.md without violating scope boundaries?

Task execution from tasks.md is managed by loading workspace artifacts, selecting the target [Txxx] task, enforcing scope.md editing rules, and updating verification.md with check results to prevent scope creep.

How does a managed change workspace keep code aligned with the spec?

A managed change workspace keeps code aligned with the spec by forcing implementation strictly against tasks.md, plan.md, and scope.md, ensuring the agent only modifies files marked as freely editable.

How to implement the next eligible task in a markdown workflow?

To implement the next eligible task, run the implement command with your change name and optional task ID, which selects the first unfinished task while honoring dependencies and required confirmations.

Can I prevent AI agents from modifying forbidden or high-risk files during implementation?

Yes, you can prevent AI agents from modifying forbidden files by enforcing scope.md edit boundaries, which stops the agent before touching high-risk surfaces and restricts changes to approved files.

When is a markdown workflow task marked complete?

A markdown workflow task is marked complete only when verification.md is updated with concrete check results that pass, or clearly explains why the checks cannot be run.

Do I need to prepare tasks.md before running task execution?

Yes, you need to prepare tasks.md before task execution because the agent requires loaded workspace artifacts to select the appropriate next eligible unfinished task and update verification status accordingly.