lsm / neokai / 25144061147

82%

push

github

Fix Codex context capacity display (#1731)

* Fix Codex context capacity metadata flow

Ensure OpenAI/Codex model aliases use local context-window metadata when provider or SDK responses report generic 200k capacities.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* Preserve Copilot model identity in metadata overlay

Keep live provider model IDs round-trippable while still applying corrected Codex context-window metadata for display and context tracking.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

* Keep context fetcher provider agnostic

Rely on preserved provider model identities for metadata matching so context fetching no longer imports Codex-specific alias logic.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

7672 of 10003 branches covered (76.7%)

Branch coverage included in aggregate %.

26 of 26 new or added lines in 3 files covered. (100.0%)

59216 of 83164 relevant lines covered (71.2%)

272.59 hits per line