Peter Steinberger
9d684e1040
docs: refresh provider usage auth refs
2026-04-04 12:40:55 +01:00
Peter Steinberger
c0d509e794
docs: refresh status cache fallback refs
2026-04-04 12:39:02 +01:00
Peter Steinberger
ac254f50e8
docs: refresh minimax usage refs
2026-04-04 12:36:18 +01:00
Vincent Koc
83c10350c6
refactor(providers): share anthropic stream helper
2026-04-04 20:35:30 +09:00
Stuart Sy
3f457cabf7
fix(status): hydrate cache usage in transcript fallback ( #59247 )
...
* fix(status): hydrate cache usage in transcript fallback
* docs(changelog): note status cache fallback fix
---------
Co-authored-by: Vincent Koc <vincentkoc@ieee.org >
2026-04-04 20:34:41 +09:00
Peter Steinberger
3100984a33
docs: refresh browser origin auth refs
2026-04-04 12:34:11 +01:00
Peter Steinberger
72847db28b
test: cover android canvas a2ui trust gate
2026-04-04 20:33:24 +09:00
Peter Steinberger
1efce6f23c
docs: refresh provider stream family docs
2026-04-04 12:32:43 +01:00
Peter Steinberger
9eb8184f36
fix: improve MiniMax coding-plan parsing ( #52349 ) (thanks @IVY-AI-gif)
2026-04-04 20:32:15 +09:00
IVY
dd9c9dac53
style: format with oxfmt
2026-04-04 20:32:15 +09:00
IVY
30de4337bf
fix: address review feedback and formatting
...
- Remove redundant name === 'MiniMax-M*' condition (already matched by startsWith)
- Use !== undefined guard instead of falsy check in deriveWindowLabelFromTimestamps
- Pass chatRemains directly to deriveWindowLabel when available
- Remove JSDoc comment style to match codebase conventions
2026-04-04 20:32:15 +09:00
IVY
efd5d5eb20
fix(usage): improve MiniMax coding-plan usage parsing for model_remains array
...
- Pick the chat model entry (MiniMax-M*) from model_remains instead of using the first BFS candidate, which could be a speech/video/image model with total_count=0.
- Derive window label from start_time/end_time timestamps when window_hours/window_minutes fields are absent; fixes the hardcoded 5h default for 4h windows.
- Include model name in plan label so users can distinguish free-tier coding-plan quota from paid API balance.
Closes #52335
2026-04-04 20:32:15 +09:00
Peter Steinberger
90af255a91
docs: refresh gemini cli usage refs
2026-04-04 12:30:55 +01:00
Peter Steinberger
65fcf7e104
fix(gateway): scope browser-origin auth throttling
2026-04-04 20:30:39 +09:00
Vincent Koc
8f7b02e567
refactor(providers): share openai stream families
2026-04-04 20:29:11 +09:00
Peter Steinberger
035a754f0f
fix: harden android a2ui trust matching
2026-04-04 20:28:08 +09:00
Peter Steinberger
1cfc10e836
docs: refresh minimax multimodal refs
2026-04-04 12:27:47 +01:00
Vincent Koc
c75f82448f
fix(google-cli): parse gemini json response and stats ( #60801 )
...
* fix(google-cli): restore gemini json reporting
* fix(google-cli): fall back to stats when usage is empty
* fix(changelog): note gemini cli cache reporting
2026-04-04 20:27:22 +09:00
Peter Steinberger
46cb493ac8
fix(sandbox): cover home credential bind audit
2026-04-04 20:27:10 +09:00
Peter Steinberger
3ec0463da9
docs: refresh minimax thinking refs
2026-04-04 12:23:33 +01:00
Peter Steinberger
3dda75894b
refactor(agents): centralize run wait helpers
2026-04-04 20:22:16 +09:00
Peter Steinberger
42778ccd46
docs: refresh provider stream family refs
2026-04-04 12:21:37 +01:00
Peter Steinberger
9615488855
fix: disable MiniMax reasoning leak ( #55809 ) (thanks @moktamd)
2026-04-04 20:21:37 +09:00
moktamd
2701e75f40
fix: disable thinking for MiniMax anthropic-messages streaming
...
MiniMax M2.7 returns reasoning_content in OpenAI-style delta chunks
({delta: {content: "", reasoning_content: "..."}}) when thinking is
active, rather than native Anthropic thinking block SSE events. Pi-ai's
Anthropic provider does not handle this format, causing the model's
internal reasoning to appear as visible chat output.
Add createMinimaxThinkingDisabledWrapper that injects
thinking: {type: "disabled"} into the outgoing payload for any MiniMax
anthropic-messages request where thinking is not already explicitly
configured, preventing the provider from generating reasoning_content
deltas during streaming.
Fixes #55739
2026-04-04 20:21:37 +09:00
Peter Steinberger
561bacd06a
fix: harden synology chat TLS helper defaults
2026-04-04 20:21:13 +09:00
Peter Steinberger
b473816afb
docs: refresh native streaming compat refs
2026-04-04 12:20:31 +01:00
Vincent Koc
bc648ac8e6
refactor(providers): add stream family hooks
2026-04-04 20:19:53 +09:00
Peter Steinberger
1037af01ad
style(agents): normalize runtime prompt formatting
2026-04-04 12:19:08 +01:00
Peter Steinberger
c70b10460c
style(auth): normalize auth choice formatting
2026-04-04 12:19:08 +01:00
Peter Steinberger
f3aad63f4e
style(providers): normalize import and wrap formatting
2026-04-04 12:19:08 +01:00
Peter Steinberger
3207c5326a
refactor: share native streaming compat helpers
2026-04-04 12:18:45 +01:00
Peter Steinberger
aaa173a4a7
docs: clarify node exec approval binding
2026-04-04 12:18:32 +01:00
Peter Steinberger
9ddfaff45f
docs: clarify node exec approval plan forwarding
2026-04-04 12:18:04 +01:00
Peter Steinberger
605f48556b
refactor(browser): share lifecycle cleanup helpers
2026-04-04 12:17:46 +01:00
Peter Steinberger
c3f415ad6e
fix: preserve node system.run approval plans
2026-04-04 20:16:53 +09:00
Peter Steinberger
f832699fd7
docs: refresh provider hook overview refs
2026-04-04 12:16:29 +01:00
Peter Steinberger
53c33f8207
fix: forward node exec approval plans
2026-04-04 20:16:19 +09:00
Peter Steinberger
62c54fdc16
docs: refresh provider replay family refs
2026-04-04 12:15:31 +01:00
Jasmine Zhang
b838ecf885
fix: add 60s timeout to MiniMax VLM fetch call
...
The VLM image analysis fetch had no timeout, causing sessions to hang
indefinitely when the MiniMax API is slow or unresponsive. Other
vision/model API calls in the codebase already use timeouts. Adds
AbortSignal.timeout(60_000) consistent with image upload workloads.
Fixes #54139
2026-04-04 20:15:13 +09:00
Peter Steinberger
39bcf695dc
fix(cron): reject unsafe custom session targets earlier
2026-04-04 20:13:39 +09:00
Peter Steinberger
00337cdde1
docs: refresh codex auth and ws refs
2026-04-04 12:11:45 +01:00
Vincent Koc
c29d4bbb86
test(providers): add family capability matrix coverage
2026-04-04 20:11:25 +09:00
Peter Steinberger
91bac7cb83
fix(usage): restore provider auth fallback
2026-04-04 12:10:45 +01:00
Peter Steinberger
6bbccb087a
docs: refresh google cached content refs
2026-04-04 12:10:29 +01:00
Peter Steinberger
49bf527fd4
docs: clarify reserved gateway method namespaces
2026-04-04 12:08:41 +01:00
Peter Steinberger
9b352ab5b0
test: isolate session status from provider runtime leak
2026-04-04 12:08:05 +01:00
Peter Steinberger
b7411ad594
refactor(cron): share descendant run quiescence wait
2026-04-04 20:07:33 +09:00
Peter Steinberger
7b6334b0f4
refactor(agents): share run wait reply helpers
2026-04-04 20:07:33 +09:00
Peter Steinberger
bbb0b574c4
refactor: centralize gateway method policy helpers
2026-04-04 20:07:18 +09:00
Vincent Koc
d766465e38
fix(google): add direct cachedContent support ( #60757 )
...
* fix(google): restore gemini cache reporting
* fix(google): split cli parsing into separate PR
* fix(google): drop remaining cli overlap
* fix(google): honor cachedContent alias precedence
2026-04-04 20:07:13 +09:00
