Spec Kit V-Model

Redefining Quality & Lifecycle Traceability

I am the creator and principal architect of the V-Model Extension Pack for Spec Kit, an open-source extension for GitHub’s Spec Kit. Since the launch of the Spec Kit Extensions website, it has ranked among the top highlighted community extensions.

The core premise of this project is simple but critical: The AI drafts. The human decides. The scripts verify. Git remembers.

This extension bridges the gap between the speed of AI-native engineering teams and the stringent traceability requirements of regulated industries. It enforces the V-Model methodology mechanically—ensuring that from a single specification, the system simultaneously generates traceable requirements, paired test plans, hazard analysis, and a deterministic traceability matrix.

Core Capabilities

• 17 Automated Commands: Spanning 4 complete V-Model levels (from requirements and acceptance testing down to module and unit design), plus cross-cutting commands like hazard analysis, impact analysis, and peer reviews.
• Built for Regulated Industries: Natively configurable for strict industry standards, including IEC 62304 (Medical Devices), ISO 26262 (Automotive functional safety), and DO-178C (Aerospace design assurance).
• Strict Gating: Enforces a rigid separation of concerns. Deterministic scripts (Bash/PowerShell) handle coverage calculations, matrix generation, and gating, while LLMs handle only the creative translation of specs.
• Massive Test Suite: Built with extreme reliability in mind, featuring over 800 automated cross-platform script tests (BATS & Pester), structural verifications, and dedicated LLM-as-judge quality evaluations.

This project sits at the exact intersection of my expertise in applied Generative AI and rigorous automotive embedded systems, demonstrating how Agentic AI can be safely and mechanically governed for mission-critical software.