What Callan Does
Callan turns Claude (or ChatGPT, or any AI chat) into a Senior Go Backend Engineer who diagnoses goroutine leaks, channel deadlocks, error-handling shape, and gRPC/REST tradeoffs from real evidence, then hands back a concrete, ready-to-execute plan.
Who It's For
Backend engineers and founders running a production Go service who need concurrency, error handling, or gRPC/REST API design decisions made correctly.
Key Capabilities
- Goroutines and channels: lifecycle ownership, the 'sender closes' rule, worker pools, and the classic leak shapes
- The sync package: Mutex/RWMutex, WaitGroup, sync.Once, and sync/atomic matched to the actual concurrency problem
- Error handling idioms: wrapping with %w, sentinel errors, custom error types, and errors.Is/errors.As instead of string comparison
- Project structure and package design: cmd/internal/pkg layout, avoiding import cycles, and constructor-based dependency injection
- Performance profiling with pprof: CPU and heap profiles, goroutine dumps, and benchmark results (ns/op, allocs/op) validating a claimed fix
- gRPC and REST API design: protobuf versioning, unary vs streaming RPCs, and picking the right protocol per service boundary
How to Use It
Download the skill file and paste the full contents into your Claude Project Instructions (or any AI chat's system prompt). Then describe your problem in plain language and get back a concrete, ready-to-execute plan - no re-explaining context every session. Works with Claude, ChatGPT, and any AI chat. Under 2 minutes to install.