build-project

Orchestrates full software project lifecycle from interview through verified final acceptance.

1|Updated Feb 6, 2026
One-click install
npx skills add https://github.com/ETdoFresh/claude-marketplace --skill build-project-etdofresh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: build-project
Source: https://github.com/ETdoFresh/claude-marketplace/tree/main/plugins/build-project/skills/build-project
Command: npx skills add https://github.com/ETdoFresh/claude-marketplace --skill build-project-etdofresh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning a rough project idea into a fully implemented, verified codebase requires coordinating planning, task breakdown, execution, and acceptance testing — a process that is error-prone and hard to keep consistent when done ad hoc. ## Core Features & Use Cases - Structured Interview & Planning: Conducts a clarifying interview, then generates a high-level plan with requirements traceability mapped to milestones. - Milestone & Task Generation with Review: Produces detailed milestone files with tasks, dependencies, and acceptance criteria, reviewed by a quality gate that loops until PASS. - Verified Execution: Executes tasks in dependency order (up to 3 in parallel) using planner, implementer, and evaluator subagents, with retry loops and a final end-to-end acceptance test. - Use Case: Describe a CLI todo app you want built; the skill interviews you about stack and scope, plans milestones, implements each task with evaluation, and only finishes when every requirement passes an automated acceptance check. ## Quick Start Ask the assistant to build your project by describing it, for example: build a command-line todo app in Python with add, list, and complete commands.

Frequently Asked Questions about build-project

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

FAQPage Schema
How do I build a complete software project from a text description?

Provide a project description as the argument, and the skill runs an interview, generates a plan and milestones, creates dependency-linked tasks, and executes them with evaluation until final acceptance passes. All artifacts are stored under projects/<slug>/.

How does automated task execution handle dependencies between tasks?

Tasks are created with TaskCreate and linked via TaskUpdate addBlockedBy based on each task's Prerequisites field. Only unblocked pending tasks are picked up, with a maximum of three executed in parallel.

What happens when a task implementation fails evaluation?

The evaluator writes a report with a PASS or FAIL result. On FAIL, the attempt counter increments and the feedback is passed back to the implementer subagent for a retry, repeating until the task passes.

Can I skip the requirements interview during project setup?

Yes. Every interview question includes a 'Skip — use your best judgment' option, and selecting it ends the interview immediately. The planner then proceeds using best-judgment assumptions for all remaining unknowns.

What are the limitations of autonomous project generation?

Each subagent runs in its own context, so all information must be passed explicitly in delegation prompts. Quality depends on the interview and plan completeness, and the process can require many retry loops for complex requirements.