Restructure prompts to move the static prefix first — one team lifted cache hit rate from 23% to 71%

AgentMarketCap·high signal

Most LLM cost dashboards ignore cache hit rate, yet it's the single biggest cost lever. Production teams report that moving a ~300-token user-context block out of the system prompt and to the start of the user turn raised cache hit rate from 23% to 71% with zero API changes and no latency risk. Treat the system prompt as a versioned artifact and assert a minimum cache hit rate in CI so it can't silently degrade as the codebase evolves.