All case studies
Developer tooling / Agentic engineering·Global·Open-source project
A governed agentic development method, shipped as open source
I built SpecForge, an open-source spec-driven lifecycle that fuses Spec Kit and AWS AI-DLC with approval gates enforced by Claude Code hooks, and run it as my forward-deployed operating system.
Governance
Hook-enforced gates
Lifecycle
Inception → Construction → Operations
Availability
Open source
The problem
Agentic coding moves fast and breaks governance. Teams need AI to propose changes while a human stays in control, with approval enforced rather than hoped for. Existing spec tools lacked hard gates.
Constraints
- Approval must be a technical barrier, not advice
- Must support brownfield reverse-engineering
- Must produce an auditable trail
What I built
- 01Fused GitHub Spec Kit's executable specs with AWS AI-DLC governance
- 02Enforced gates with PreToolUse hooks that block edits until specs/plans pass review
- 03Added a Ralph autonomous loop that respects the approved plan
- 04Shipped it as a Claude Code plugin with a committed audit log
Stack
Claude CodeSpec KitAWS AI-DLCTypeScriptHooks
Engagement
Want a similar outcome?
Every engagement is scoped fresh. Book a discovery call to see if there's a fit, or send a brief in writing.
AI engineering engagements start at $18k.
Agent platform builds, MCP servers, and production AI integration. Day rate from $1,800 for embedded work. Most engagements run 4 to 12 weeks.