blueprint

Creates a multi-phase construction plan from a single objective for software projects.

2|Updated Mar 20, 2026
One-click install
npx skills add https://github.com/multiplex-ai/muggle-ai-teams --skill blueprint-multiplex-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: blueprint
Source: https://github.com/multiplex-ai/muggle-ai-teams/tree/main/skills/blueprint
Command: npx skills add https://github.com/multiplex-ai/muggle-ai-teams --skill blueprint-multiplex-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Turn a single objective into a step-by-step construction plan for multi-session, multi-agent engineering projects. Each step has a self-contained context brief so a fresh agent can execute it cold. Includes adversarial review gate, dependency graph, parallel step detection, anti-pattern catalog, and plan mutation protocol.

Core Features & Use Cases

  • Five-phase pipeline: Research, Design, Draft, Review, Register, with cold-start context for each step.
  • Adversarial review gate to ensure completeness and correctness.
  • Dependency graph and parallel/serial scheduling to optimize multi-agent collaboration.
  • Plan mutation protocols to insert, split, or reorder steps with audit trails.
  • Use cases include breaking large features into PR-sized steps and coordinating cross-team work across sessions.

Quick Start

Describe your objective to Blueprint using the /blueprint command to generate a multi-step construction plan.

Frequently Asked Questions about blueprint

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

FAQPage Schema
How do I break down a large software feature into PR-sized steps for multi-agent execution?

To break down large features, you can generate a multi-phase construction plan that splits objectives into PR-sized steps. Each step includes a self-contained context brief, allowing fresh agents to execute tasks cold without prior session memory.

What is a dependency graph in multi-agent project planning and how does it optimize parallel workstreams?

A dependency graph in multi-agent project planning maps step relationships to schedule parallel and serial execution. It optimizes multi-agent collaboration by identifying which steps can run simultaneously and which require prior completion.

How do I ensure plan completeness and correctness when coordinating cross-team engineering work?

To ensure plan completeness, an adversarial review gate evaluates the drafted steps before registration. This review phase checks the multi-phase construction plan for correctness, validating that all step contexts are self-contained and dependencies are accurately mapped.

Can I insert, split, or reorder steps in a multi-phase construction plan without losing traceability?

Yes, you can mutate a multi-phase construction plan using dedicated plan mutation protocols. These protocols allow you to insert, split, or reorder steps while maintaining audit trails, ensuring dependency graphs remain accurate across sessions.

What's the best way to turn a single objective into a structured multi-phase construction plan?

The best way to turn a single objective into a structured plan is through a five-phase pipeline: research, design, draft, review, and register. This process enforces cold-start execution contexts and adversarial review before finalizing the dependency graph.

Do I need prior session memory for an agent to execute a step in a multi-session engineering plan?

No, agents do not need prior session memory to execute steps. Each step in the construction plan provides a self-contained cold-start context brief, ensuring a fresh agent can perform the task independently without prior workflow knowledge.