Building dependable software, purposeful tools, and practical infrastructure.
I build practical software for real repositories: coding-agent workflows, developer automation, secure admin systems, and production-grade web platforms.
|
|
|
AgentContextKit An offline-first developer toolkit that helps make repositories safer and easier for AI coding agents to understand and work with. |
ACKit Spec Kit Bridge Community bridge between GitHub Spec Kit and ACKit that maps spec-driven artifacts into deterministic task state, evidence, state-bound verification, completion gates, checkpoints and handoffs. |
   
|
  
|
| Repository → · npm → · Documentation → | Repository → · npm → · Documentation → |
AgentContextKit focuses on:
- Repository and instruction analysis.
- Agent-ready context and workflow preparation.
- Task-first and docs-first development workflows.
- Security, secret, privacy, and repository-hygiene checks.
- CI-friendly reports and automation-ready outputs.
Project details, installation, commands, architecture, development status, and release information live in the AgentContextKit repository.
| Area | Direction |
|---|---|
| Engineering workflows | Safer context, agent workflows, and developer automation |
| Developer tooling | Practical tools for real repositories and repeatable engineering workflows |
| Production web systems | ASP.NET Core MVC, admin panels, RBAC, audit logs, SEO, deployment |
| Operations | GitHub Actions, Docker, Linux, Cloudflare, MSSQL, server workflows |
TypeScript Node.js .NET ASP.NET Core C# MSSQL EF Core GitHub Actions Docker Linux Cloudflare Git
Understand the system → Define the change → Implement deliberately → Verify the result.
| Principle | Meaning |
|---|---|
| Docs first | Changes start with context, scope, and documentation. |
| Task first | Acceptance criteria and rollback notes before implementation. |
| Security-aware | Secrets, permissions, audit trails, and safe defaults matter. |
| Production-ready | Built for real environments, not throwaway demos. |
| Maintainable | Prefer readable, testable, predictable technology. |
Open an issue or discussion on any public repository.