open-code-review/internal/llm
kite 08bb28ced5 feat(llm): enable Anthropic prompt caching and fix token accounting
Add cache_control ephemeral breakpoints on system prompt and tool
definitions to activate Anthropic prompt caching for multi-turn
agent loops. Fix token double-counting by removing the redundant
totalTokensUsed accumulator and computing it from sub-items. Surface
cache read/write statistics in both text and JSON summary output.
2026-06-06 19:26:34 +08:00
..
bpe_data feat: init 2026-05-20 22:03:52 +08:00
client.go feat(llm): enable Anthropic prompt caching and fix token accounting 2026-06-06 19:26:34 +08:00
client_test.go feat(llm): enable Anthropic prompt caching and fix token accounting 2026-06-06 19:26:34 +08:00
embedded_loader.go feat: init 2026-05-20 22:03:52 +08:00
resolver.go feat(resolver): strip model name suffix and prioritize config file over env vars 2026-05-26 17:58:03 +08:00
resolver_test.go feat(resolver): strip model name suffix and prioritize config file over env vars 2026-05-26 17:58:03 +08:00
usage_resolver.go style: apply go fmt formatting 2026-06-03 17:34:28 +08:00