{"avoid_when":"一次性脚本、不想引入优化循环","content_hash":"fa32785b569729141b58d8481d28e1a1d8ff625e7981363a1f1bf7a1e52510a9","domain":"ai-agents","evidence_urls":["https://github.com/stanfordnlp/dspy"],"id":"dspy","language":"Python","license":"","name":"DSPy","niche":"programmatic prompt optimization","repo":"https://github.com/stanfordnlp/dspy","source_updated_at":null,"status":"active","summary":"斯坦福声明式提示与程序优化框架，把 prompt 当可优化参数。","tags":["orchestration","prompt"],"use_when":"要用数据驱动优化提示/程序而非手写死 prompt","verification_status":"unverified","verified_at":null}