skills.rest
baobao0303
baobao0303COMMUNITY
Explore all skills by this author →
WebsiteGitHub

harness-agent-dev

Execute approved story implementation with test-first red/green/refactor discipline.

1|Updated May 23, 2026
Search Tags:#TDD#AI coding agent#project context#story execution#code implementation#architectural decisions#test-first development
One-click install
npx skills add https://github.com/baobao0303/harness --skill harness-agent-dev
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-agent-dev
Source: https://github.com/baobao0303/harness/tree/main/.agents/skills/harness-agent-dev
Command: npx skills add https://github.com/baobao0303/harness --skill harness-agent-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python3.

What problem does it solve?

It turns an approved implementation request into a disciplined story-execution workflow, ensuring the AI focuses on the right work, follows acceptance criteria, and ships test-verified changes.

Core Features & Use Cases

  • Story execution with a test-first loop: Runs the approved story workflow using red/green/refactor discipline and focuses on passing acceptance criteria.
  • Deterministic activation pipeline: Resolves customization via a resolver script (with a file-based structural merge fallback), runs prepend/append activation steps, and then dispatches to the appropriate dev sub-skill.
  • Evidence-ready context loading: Loads persistent facts (including project-context files) and project config from the Harness config to align code changes with required validation artifacts.

Quick Start

Use the harness-agent-dev skill to implement an approved story by responding with: "Amelia, implement the next story from the Harness backlog for me."

Frequently Asked Questions about harness-agent-dev

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

FAQPage Schema
How do I use an AI coding agent for test-first development?▼

Test-first development with an AI coding agent uses a red/green/refactor loop to implement approved stories. The agent focuses on passing acceptance criteria and shipping test-verified code changes.

How do I execute approved project stories with verified code delivery?▼

Executing approved stories requires loading project context and persistent facts to align code changes with validation artifacts. The agent dispatches to sub-skills to implement work based on acceptance criteria.

Do I need Python3 to run story execution workflows?▼

Python3 is required to run the story execution workflow. The deterministic activation pipeline uses a resolver script to handle customization and dispatch to the appropriate development sub-skill.

What is the best way to align AI code implementation with architectural decisions?▼

Aligning AI code implementation with architectural decisions relies on loading persistent facts and project config from the Harness configuration. This evidence-ready context ensures changes meet required validation artifacts.

Can I use story execution for sprint planning and code review?▼

Story execution supports sprint planning, code review, testing, and retrospective menus. The developer persona handles story preparation, quick dev clarification, and investigation menus within the workflow.

Why does my AI coding agent need a deterministic activation pipeline?▼

A deterministic activation pipeline ensures consistent agent behavior by resolving customization via a resolver script with a file-based structural merge fallback. It runs prepend and append steps before dispatching to sub-skills.

Related Skills

Explore More Software Engineering Skills→
enigmaicon-engenigmaicon-eng
1

bmad-agent-dev

Implement approved enterprise stories using a test-first red/green/refactor workflow.

Community
Intermediate
CaptainCrouton89CaptainCrouton89
497

Testing Code

Automate software tests to verify functionality against acceptance criteria.

Community
Basic
cassioalexandrecassioalexandre
2

bmad-agent-sm

Convert sprint goals and story requests into structured sprint plans and developer-ready stories.

Community
Intermediate
xenitV1xenitV1
230

tdd-mastery

Enforce test-first discipline with failing tests before production code.

Community
Basic