Devika Verma f589013e84 Carry cache_read/cache_write through Verdict, stats, and the eval report ...

Auto-caching providers (OpenAI/Together/Gemini) serve most of the
reviewer's 1,400-token prefix from cache and report the split; dropping
it made calls read as '16 in' - present-tense misreporting, not a v2
optimization (owner call 2026-08-17). Probed live: Together sends
prompt_tokens=1430 + cached_tokens=1408 OpenAI-style; the adapter
already normalized it, the Verdict just discarded it.

Verdict + reviewer.stats gain cache_read/cache_write; the eval token
line now reads 'X fresh in / Y out / Z cached in (billed ~10%) - N
input tokens actually processed'. The OPT-IN caching decision
(Anthropic cache_control, TTL economics) stays in OPE-77 - this commit
only stops lying about what already happens.

2026-08-17 22:41:03 +05:30

..

eval_reviewer.py

Carry cache_read/cache_write through Verdict, stats, and the eval report

2026-08-17 22:41:03 +05:30