- Add missing claude-sonnet-4-7/4.7 and claude-haiku-4-7/4.7 mappings;
previously claude-sonnet-4.7 was substring-matched by the bare
"claude-sonnet-4" key and silently downgraded to claude-sonnet-4.
- Introduce modelMapping.boundary flag and modelKeyMatches() helper.
Bare digit-ending keys (like claude-sonnet-4) now require the next
character to NOT be a digit, dot, or dash-digit, so future versions
(4.8, 5.x) also pass through without silent downgrade.
- Add 8 regression tests in TestParseModelAndThinkingNoSilentDowngrade
covering the 4.7 family, hypothetical 4.8, Bedrock-style names, and
thinking-suffix variants.
* fix: stabilize multimodal image compatibility across OpenCode flows
Advertise vision-capable metadata in /v1/models and make model matching deterministic so OpenCode does not downgrade image support or route 4.6 models incorrectly. Expand request translation to accept OpenCode/OpenAI attachment shapes, sanitize [Image N] placeholders safely, keep image-only follow-up turns non-empty, and improve token accounting so base64 image bytes no longer inflate prompt token usage and trigger premature compaction.
* fix: deduplicate thinking streams and trim injected prompt noise
* fix: align /v1/messages thinking blocks and message_start usage
* fix: reduce repetitive thinking across tool turns
Select a single reasoning stream source, prevent chunk replay, and preserve structured tool-loop context so the model keeps continuity instead of re-planning each turn.
* fix: unify token counting on existing API endpoints
Compute usage deterministically on /v1/messages and /v1/chat/completions even when upstream omits tokenUsage.
- remove roo-only token path and keep behavior on existing endpoints
- add proxy/token_estimator.go with shared Claude/OpenAI estimators (input/system/messages/tools + output/thinking/tool calls)
- wire stream/non-stream handlers to use estimator-derived input/output usage
- update /v1/messages/count_tokens to reuse the same estimator
- keep robust upstream usage parsing/normalization in proxy/kiro.go while dropping parser-level estimate fallback
Why: direct upstream tests show metering/context events frequently arrive without tokenUsage in this environment; this made usage zero or inconsistent. Local deterministic accounting keeps reported usage stable and explicit.
