Summary
Create an ARCHITECTURE.md documenting the project's design and data flow. Foundation reviewers and new contributors need this to understand the codebase quickly.
Suggested sections
- Overview — what gpu-mcp-server does and where it fits
- Data flow — NVML → Collector → Handler → MCP SDK → Transport → Agent
- Key interfaces —
Collector (gpu.go), Handler (server.go)
- Build tags —
nvml.go (cgo+linux), stub.go (everything else), mock.go (tests)
- Transport modes — stdio (default), HTTP (optional)
- Deployment — standalone binary, Docker, Kubernetes DaemonSet (Helm)
- MIG support — how MIG instances are detected and reported
Reference
See keda-gpu-scaler's ARCHITECTURE.md for the expected level of detail.
Acceptance criteria
Difficulty: Beginner-Intermediate — documentation, requires reading the code
Summary
Create an
ARCHITECTURE.mddocumenting the project's design and data flow. Foundation reviewers and new contributors need this to understand the codebase quickly.Suggested sections
Collector(gpu.go),Handler(server.go)nvml.go(cgo+linux),stub.go(everything else),mock.go(tests)Reference
See keda-gpu-scaler's
ARCHITECTURE.mdfor the expected level of detail.Acceptance criteria
ARCHITECTURE.mdat repo rootDifficulty: Beginner-Intermediate — documentation, requires reading the code