Your agent starts a run with cache_read_input_tokens at 40K and climbing. Twelve tool calls later, reads drop to zero and cache_creation_input_tokens jumps to the full conversation length — on every single turn. Nothing in your prompt changed.
Source: [Dev.to](https://dev.to/ji_ai/claude-prompt-caching-why-agent-loops-miss-the-20-block-lookback-d36)