blueprint

Generate ordered executable step plans with dependency graphs and review gates.

1|Updated May 12, 2026
One-click install
npx skills add https://github.com/Manvendra08/TradingBot --skill blueprint-manvendra08
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: blueprint
Source: https://github.com/Manvendra08/TradingBot/tree/main/_agent/skills/blueprint
Command: npx skills add https://github.com/Manvendra08/TradingBot --skill blueprint-manvendra08

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the risk of context loss and rework when breaking large, multi-session engineering tasks (such as multi-PR feature builds, cross-session refactors, or database migrations) into executable steps, ensuring fresh agents can pick up work without prior context.

Core Features & Use Cases

  • Step-by-step construction plans: Breaks complex objectives into 3–12 one-PR-sized steps with clear dependency ordering, parallel step detection, and self-contained context briefs for cold-start execution.
  • Adversarial review gate: Every plan is reviewed by a high-capability sub-agent against an anti-pattern checklist to catch completeness, dependency, and design errors before execution.
  • Use case: Coordinate parallel workstreams across multiple sub-agents for tasks like extracting LLM providers into a plugin system or migrating a database to a new backend.

Quick Start

Use the blueprint skill to turn your objective of refactoring the TradingBot's signal detection pipeline into a step-by-step plan with parallel workstreams and review gates.

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 code refactoring project into multi-session engineering plans?

Multi-session engineering plans break large refactoring objectives into 3–12 one-PR-sized steps with clear dependency ordering and self-contained context briefs. This enables fresh coding agents to execute steps without prior context, eliminating rework risks.

What is the best way to coordinate parallel workstreams across multiple sub-agents?

Coordinating parallel workstreams requires detecting parallelizable steps within an engineering roadmap and assigning them to multiple sub-agents. This approach generates formal plans with dependency graphs, enabling concurrent execution of independent tasks like plugin extraction.

How does an adversarial review gate prevent errors in multi-PR feature development?

An adversarial review gate checks every generated step plan against an anti-pattern checklist using a high-capability sub-agent. It catches completeness, dependency, and design errors before execution, ensuring the multi-PR feature development plan is robust.

Can I execute a cold-start database migration plan without prior session context?

Cold-start database migration execution is possible when plans contain self-contained context briefs and formal mutation protocols. Any coding agent can pick up the executable step plan and resume work without needing prior session history.

When should I avoid using automated engineering roadmaps for code refactoring?

Automated engineering roadmaps are unnecessary for small, single-session code refactoring tasks. They are designed for complex, multi-session objectives where context loss and rework risks are high, such as cross-session refactors or multi-PR feature builds.