llm-cache-boundary-design

Designs prompt-prefix, response, semantic-response, and document-context caching for LLM workloads, including Claude native prompt caching and cache-augmented generation, with invalidation and miss-path planning. Use when the user asks to "cache prompts", "reduce LLM costs", "speed up repeated prompts", "implement response caching", or complains about stale cached answers or cache misses. Not for model training or fine-tuning, for vector-store selection in RAG, or for changing provider account settings.

Version

v1.0.0

Size

8 KB

Updated

2026-10-02

Your client verifies the signature before reading. This explicit manual read is available on every plan. Free automatically loads five chosen skills; paid plans automatically load relevant skills across the library.