ArchGuardian

ArchGuardian provides deep visibility into software architecture, identifies risks, fixes broken dependencies, updates libraries, and prevents technical debt from accumulating.

Comment

Inspiration

While working on another gargantuan project (knirv.com), it occurred to me that I needed help with this. Aside from trying to recruit as many bros as possible to pick up on the vibe coding wave, since my budget cannot afford hiring an actual coder at this time... I figured, why not look into an agentic solution? I tried utilizing the already existing CodeLogic platform, which led to an extreme crash of my entire Linux system. After losing a day or so of work rebuilding my entire Linux desktop, I went as far as deploying the CodeLogic platform in a virtual machine (what I should have done in the first place)... Yet and still to no avail, the process was simply too complex and resource-intensive. I figured, why not build this myself with third-party inference providers? So, I did!

What it does

ArchGuardian performs three core functions flawlessly:

1. Understands Your System

Creates a real-time knowledge graph showing how every component connects—code, APIs, databases, services, and data flows across your entire stack.

2. Diagnoses Hidden Risks

Surfaces technical debt, security vulnerabilities, obsolete code, and dangerous dependencies before they become critical failures.

3. Automates the Fixes

AI generates and executes fixes—updating libraries, removing dead code, patching vulnerabilities, and preventing new debt from accumulating.

How we built it

I simply copied the features listed on the CodeLogic site into Claude and requested that it give me a Golang program, then downloaded the initial main.go file onto my system and built the system out from there. The result is far more complex than what we started with from Claude, but the journey has been rewarding in so many ways, and I'm glad we went in this direction.

Challenges we ran into

Frontend-Backend Disconnect: Significant gaps between what frontend expected and backend provided Data Flow Issues: Scanner/diagnoser generated data but APIs returned mock data Testing Coverage: Embedding service failures blocking tests Multi-Project Support: Partially implemented but not fully functional

Accomplishments that we're proud of

✅ Strong Foundation: Well-structured Go backend with modular design

  • ✅ AI Integration: Multi-provider AI orchestration (Cerebras, Gemini, Anthropic, OpenAI, DeepSeek)
  • ✅ Data Persistence: Chromem-go vector database integrated and functional

  • ✅ Authentication: GitHub OAuth implementation complete

    What we learned

    We learned can be a part of the CVE reporting structure as detailed in our vulnerability disclosure:

  • Disclosure: Vulnerability Disclosure

  • Documentation: Full documentation

  • Issues: GitHub Issues

What's next for ArchGuardian

Upcoming Features

  • [ ] Multi-language Support: Python, JavaScript, TypeScript, Java
  • [ ] CI/CD Integration: GitHub Actions, GitLab CI, Jenkins plugins
  • [ ] Slack/Discord Notifications: Real-time alerts
  • [ ] Custom Rule Engine: Define your own code quality rules
  • [ ] Team Collaboration: Multi-user support with role-based access
  • [ ] Cloud Deployment: Docker, Kubernetes, AWS/GCP/Azure templates
  • [ ] Advanced Analytics: ML-powered trend prediction
  • [ ] IDE Plugins: VSCode, IntelliJ, Vim extensions
Share this project:

Updates