go-temporal
Installation
SKILL.md
This skill guides creation of production-ready Temporal applications following Go idioms and best practices. Build durable, fault-tolerant workflows with proper error handling, observability, and testing.
Core Mental Model
Temporal separates what happens from when it happens:
- Workflows: Deterministic orchestrators that survive failures. They coordinate, make decisions, and maintain state. Think of them as reliable state machines that never lose progress.
- Activities: Unreliable operations (I/O, APIs, databases) wrapped with automatic retries. They do the actual work but may fail transiently.
- Workers: Stateless executors polling task queues. Scale horizontally by adding more workers.
Key insight: Workflows replay from history on restart. Any non-deterministic operation (time, random, I/O) breaks replay. Use workflow.* APIs for determinism.