Kiro's upstream model is trained to identify and resist
--- SYSTEM PROMPT --- marker blocks as injection attempts, causing it
to actively reject the user's system prompt and self-correct its
identity. Switch the Claude path to the same plain-prepend approach
already used by the OpenAI path: system content is joined directly
before the user message without any marker, matching natural context.
The sanitizer (reSysPromptBlock) still strips the old marker format
from conversation history until existing contamination clears out.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When the proxy's own --- SYSTEM PROMPT --- wrapper or Claude Code's
<system-reminder> blocks appear in conversation history (e.g. echoed
back by Kiro and included in the next request), strip them from user
and assistant message content before building the Kiro payload.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Brought in 9 upstream commits:
- 221348b thinking routing: ClaudeRequest.Thinking + Signature + includeEmptyThinkingBlock
- 0203357 + 31aa6aa accurate input_tokens via contextUsageEvent
- 404e242 + 50f1a7e outbound proxy (socks5/http) + UI
- 940dc78 version bump to 1.0.6
- 3 CI workflow changes
Strategy: took upstream base for the 4 conflicting files, then re-applied
our local changes on top:
- config.go: InvalidModelRetries field + GetInvalidModelRetries/UpdateInvalidModelRetries
- kiro.go: AmazonQ origin CLI->AI_EDITOR, attempt-level retry loop for
INVALID_MODEL_ID, detailed log.Printf (account/model/attempt/elapsed),
log import; adopted upstream's kiroHttpStore atomic pointer for Do()
- handler.go: /admin/api/general GET/POST + apiGetGeneralConfig +
apiUpdateGeneralConfig
- web/index.html: General Settings card (invalid-model-retries),
CN/EN i18n, loadGeneralConfig/saveGeneralConfig, call from initSettings
Build + full test suite green on Go 1.24.3.
- 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.
- Add claude-sonnet-4.6 (dot and dash variants) to modelMap in translator.go
- Add claude-sonnet-4.6 and claude-opus-4.6 (plus -thinking variants) to the
static fallback model list in handler.go
- Realign existing opus-4.6 entries for consistency
- Multi-account pool with round-robin load balancing
- Auto token refresh for IAM IdC and Social auth
- Streaming support (SSE)
- Web admin panel with account management
- Docker support with GitHub Actions CI/CD
- Machine ID management per account
- Usage tracking (requests, tokens, credits)
