df123a1979
test(mistral): wait on delivered transcripts after reconnect ( #124935 )
...
* test(mistral): wait on delivered transcripts after reconnect
* test(nvidia): assert the remaining catalog fetch budget as a bound
2026-08-16 17:45:17 -07:00
8f46e31994
fix(nvidia): hide retired models while preserving live vendor discovery ( #117620 )
...
* fix(model-catalog): preserve manifest lifecycle metadata on refresh
* fix(nvidia): simplify Qwen model lifecycle
* test(nvidia): avoid map-spread lint
* docs(nvidia): clarify retired Qwen models
* fix(models): preserve NVIDIA lifecycle states
* fix(models): keep NVIDIA lifecycle plugin-owned
* fix(nvidia): retain fresh featured catalog models
---------
Co-authored-by: clawsweeper <274271284+clawsweeper[bot]@users.noreply.github.com>
Co-authored-by: Peter Steinberger <steipete@macos.shared >
Co-authored-by: Peter Steinberger <steipete@gmail.com >
2026-08-02 04:40:17 -07:00
93a0fdb9b8
fix(release): align extension validation contracts ( #113919 )
2026-07-25 17:07:51 -07:00
f703d803cc
chore(models): curate all provider model catalogs to current-generation lineups ( #113681 )
...
* chore(models): fleet-wide provider catalog curation
* fix(models): sync provider runtime catalogs and tests with curated manifests
* test(models): align venice lifecycle assertions and copilot auth default with curated catalogs
* test(models): align catalog lifecycle contract checks
2026-07-25 07:21:17 -07:00
370b80b74d
refactor(deadcode): tighten provider extension roots ( #108497 )
2026-07-15 16:23:47 -07:00
47ec6237ff
fix(nvidia): sync recommended models with featured catalog ( #99242 )
...
* fix(nvidia): refresh bundled featured models
* fix(nvidia): match featured model feed exactly
* fix(nvidia): remove transient feed detail
* fix(nvidia): correct super context window
* fix(nvidia): align auth choice expectation
* fix(nvidia): hide deprecated models from selection
Refresh the bundled featured catalog while retaining deprecated model metadata for exact-reference compatibility. Route NVIDIA onboarding and catalog picker surfaces through selectable-only provider builders.
* fix(nvidia): simplify featured model table
* fix(nvidia): restore shipped compatibility rows
2026-07-09 13:09:38 -07:00
cc76e06e7a
test(plugins): align release fixtures with runtime contracts
2026-07-01 18:34:53 -07:00
fe43343af0
fix(nvidia): use Nemotron Super 1M context ( #98726 )
...
* fix(nvidia): correct Nemotron Super context
2026-07-01 11:46:06 -07:00
57e0bdaabe
feat: add live provider model catalog helper
...
Summary:
- Add a shared live provider catalog runtime for SDK-backed providers.
- Route OpenAI, xAI, OpenCode Go, Chutes, DeepInfra, Venice, NVIDIA, and Vercel AI Gateway live model discovery through the shared helper.
- Remove duplicated provider-local live catalog caching and harden auth marker stripping, empty live-result retries, and OpenAI custom-base-url handling.
Verification:
- node scripts/run-vitest.mjs extensions/openai/openai-provider.test.ts src/plugin-sdk/provider-catalog-live-runtime.test.ts src/commands/models/list.source-plan.test.ts extensions/opencode-go/index.test.ts extensions/nvidia/provider-catalog.test.ts
- pnpm plugin-sdk:api:check
- pnpm lint --threads=8
- pnpm run lint:extensions:bundled
- pnpm run test:extensions:package-boundary:compile
- pnpm check:import-cycles
- pnpm exec oxfmt --check extensions/openai/openai-provider.ts extensions/openai/openai-provider.test.ts
- git diff --check origin/main...HEAD
- autoreview clean: no accepted/actionable findings reported
- AWS Crabbox focused remote proof: run_364680d1bff8 / cbx_2456fffafe01
- Earlier same-PR AWS Crabbox live proof: run_1f05ccab368e / cbx_7375c79fcf9b
Known proof gap:
- Final current-code true live-provider smoke was blocked by Crabbox secret hydration, documented in the PR proof comment.
2026-06-07 14:16:00 -07:00
4fa5092cdc
docs: document small extension sources
2026-06-04 21:02:07 -04:00
829847292e
feat(nvidia): default to nemotron ultra
...
Add NVIDIA Nemotron 3 Ultra to the bundled NVIDIA provider catalog and make it the bundled fallback default. Keep successful NVIDIA featured-model feeds authoritative, while treating the first live feed row as the setup default option. Update NVIDIA provider docs and focused provider/model-picker tests for the new Ultra behavior.
Verification:
- node scripts/run-vitest.mjs src/commands/model-picker.test.ts extensions/nvidia/provider-catalog.test.ts extensions/nvidia/index.test.ts extensions/nvidia/onboard.test.ts
- pnpm exec oxfmt --check src/flows/model-picker.ts src/commands/model-picker.test.ts
- pnpm format:docs:check
- pnpm docs:check-mdx
- git diff --check upstream/main...HEAD
- .agents/skills/autoreview/scripts/autoreview --mode branch --base upstream/main --parallel-tests "node scripts/run-vitest.mjs src/commands/model-picker.test.ts extensions/nvidia/provider-catalog.test.ts extensions/nvidia/index.test.ts extensions/nvidia/onboard.test.ts"
2026-06-04 20:13:06 +01:00
60f8e18372
fix(nvidia): bound featured model cache expiry
2026-05-30 12:34:53 -04:00
6fd4aa8a27
fix(nvidia): load featured model catalog ( #80775 )
...
* fix(nvidia): load featured model catalog
Co-authored-by: CaptainTimon <CaptainTimon@users.noreply.github.com >
* fix(nvidia): widen catalog fetch timeout
* fix(nvidia): cover catalog registration
* fix(picker): include provider catalog loader
* fix(nvidia): guard featured catalog fetch
* fix(nvidia): sync bundled catalog with live API
Replace minimaxai/minimax-m2.5 (MiniMax M2.5) with minimaxai/minimax-m2.7 (Minimax M2.7) and z-ai/glm5 (GLM-5) with z-ai/glm-5.1 (GLM 5.1) in the bundled fallback catalog to match NVIDIA's public featured-models endpoint.
Update docs table and all extension test expectations.
* fix(nvidia): retain shipped catalog refs
* fix(picker): keep alias catalog rows
* fix(nvidia): restore live catalog priority
---------
Co-authored-by: CaptainTimon <CaptainTimon@users.noreply.github.com >
2026-05-28 12:59:55 -07:00
05fd67f822
test: clarify nvidia provider assertions
2026-05-08 11:51:58 +01:00
dcd665cd05
fix(nvidia): align NIM provider metadata
...
Persist the NVIDIA_API_KEY marker in generated catalog output and mark bundled NVIDIA Chat Completions models as string-content compatible.\n\nFixes #73013.\nFixes #50107.\nRefs #73014 .
2026-04-28 11:30:57 -07:00
d43cc470c6
refactor(nvidia-endpoints): updated language & default models ( #59866 )
...
* fix(nvidia-endpoints): updated language & default models
* fix(nvidia-endpoints): updated link for api key
* fix(nvidia-endpoints): removed unused const
* fix(nvidia-endpoints): edited max tokens
* fix(nvidia-endpoints): fixed typo
---------
Co-authored-by: Devin Robison <drobison00@users.noreply.github.com >
2026-04-07 08:47:29 -06:00
adc329b26b
test: dedupe extension-owned coverage ( #58554 )
...
* test: dedupe extension-owned coverage
* test: remove duplicate coverage files
* test: move helper coverage into extensions
* test: trim duplicate helper assertions
* test: remove cloudflare helper import from agent test
* test: align stale expectations with current main
2026-03-31 15:18:29 -07:00
Peter Steinberger