bulletproof

Coordinate a 12-stage adaptive development workflow from research to deployment.

146|24|Updated Mar 21, 2026
One-click install
npx skills add https://github.com/artemiimillier/bulletproof --skill bulletproof
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bulletproof
Source: https://github.com/artemiimillier/bulletproof/tree/main
Command: npx skills add https://github.com/artemiimillier/bulletproof --skill bulletproof

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Bulletproof provides a disciplined, repeatable process to manage AI-driven software work, preventing chaotic coding and unfinished tasks by enforcing a 12-stage workflow from research to deployment.

Core Features & Use Cases

  • End-to-end 12-stage workflow: research, spec, plan, phased implementation, self-audit, verification, impact analysis, integration checks, code review, security scan, fixes, and deploy.
  • Context management and gates: fresh-context handoffs, 40% context window guardrails, and deterministic progress checks to maintain quality across stages.
  • Anti-rationalization and guardrails: detects incomplete work, requires justification and evidence before moving forward, and prevents scope creep.
  • Use case coverage: ideal for feature development, refactors, complex bug fixes, or architectural changes in software projects.

Quick Start

Run the bulletproof skill via Claude Code to activate the 12-stage workflow for your non-trivial coding task.

Frequently Asked Questions about bulletproof

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

FAQPage Schema
How do I prevent AI agents from leaving software tasks unfinished or producing chaotic code?

To prevent chaotic coding and unfinished tasks, you can enforce a rigorous 12-stage AI development workflow with strict stage gates, anti-rationalization checks, and fresh-context handoffs. This ensures verifiable, reliable delivery from research to deployment.

What is a fresh-context handoff in AI-assisted software development?

A fresh-context handoff is a process mechanism in AI development workflows where task context is reset or managed between stages. It works alongside 40% context window guardrails to maintain reasoning quality and prevent context degradation during complex coding tasks.

How do I manage complex feature development and refactoring using AI agents?

You can manage complex feature development and refactors by applying an adaptive 12-stage workflow that guides AI agents step-by-step through research, spec, planning, phased implementation, self-audit, verification, impact analysis, and deployment.

Does this 12-stage AI workflow work for complex bug fixes and architectural changes?

Yes, this 12-stage workflow is explicitly designed for complex bug fixes and architectural changes across software projects. It enforces deterministic progress checks, security scans, and integration checks to verify reliable delivery before moving forward.

What's the best way to stop AI scope creep during automated code generation?

The best way to stop AI scope creep is to implement anti-rationalization guardrails that detect incomplete work and require justification and evidence before advancing. Strict stage gates ensure the AI agent remains within the defined task scope.

When should I not use a rigid 12-stage development workflow for AI coding?

You should not use a rigid 12-stage workflow for trivial or minor coding adjustments, as the strict stage gates, fresh-context handoffs, and mandatory verification checks are designed specifically for non-trivial software tasks and complex architectural changes.