f5a50db569
fix(microsoft-foundry): reject malformed endpoints ( #104796 )
2026-07-12 08:23:44 +01:00
7bf80dc2c6
chore(tooling): enforce formatting and refresh TypeScript checks ( #104239 )
...
* chore(tooling): enforce current formatter and refresh checks
* chore(tooling): keep release changelog formatter-owned
* chore(tooling): retain compatible Node type surface
* ci: enforce formatting for docs-only changes
* ci: isolate docs formatter check
* chore(tooling): apply updated lint and format rules
* chore(tooling): satisfy updated switch lint
* style(ui): apply Linux formatter layout
* test(doctor): match quiet local audio contribution
* test(doctor): assert quiet output only
* test(doctor): follow restored information contract
2026-07-11 01:09:51 -07:00
c335ecdb73
fix(microsoft-foundry): keep Azure CLI error truncation UTF-16 safe ( #102604 )
...
* fix(microsoft-foundry): keep Azure CLI error truncation UTF-16 safe
Replace `.slice(0, 300)` with `truncateUtf16Safe(normalized, 300)` in
summarizeAzErrorMessage to prevent surrogate pair corruption.
Ref. lsr911 pattern — mechanical substitution, no behavior change.
* test(microsoft-foundry): cover UTF-16-safe CLI errors
* test(microsoft-foundry): assert complete CLI error
---------
Co-authored-by: Peter Steinberger <steipete@gmail.com >
2026-07-09 11:49:35 +01:00
abec88cfab
fix(auth): apply runtime request overrides everywhere
2026-06-11 11:11:22 +09:00
08655fb02b
fix(foundry): bind auth and thinking contracts
2026-06-11 11:11:22 +09:00
76ce9d6d22
fix(foundry): expose Claude 4.6 max effort
2026-06-11 11:11:22 +09:00
78d621855d
fix(foundry): label Anthropic Messages onboarding
2026-06-11 11:11:22 +09:00
819fc0037a
fix(foundry): align Claude thinking contracts
2026-06-11 11:11:22 +09:00
bebe96b402
test(foundry): type auth lookup
2026-06-11 11:11:22 +09:00
818c3c276e
fix(foundry): preserve canonical thinking identity
2026-06-11 11:11:22 +09:00
9ffb5407f7
fix(providers): reconcile Claude profile routing
2026-06-11 11:11:21 +09:00
4457ae3e35
fix(providers): preserve Claude output limits
2026-06-11 11:11:21 +09:00
55384bc236
fix(foundry): clamp Mythos Preview thinking
2026-06-11 11:11:21 +09:00
9db8601235
fix(foundry): avoid stale setup metadata
2026-06-11 11:11:21 +09:00
2031603542
fix(providers): preserve Mythos max routing
2026-06-11 11:11:21 +09:00
4afe616f22
fix(providers): apply auth patch deletions
2026-06-11 11:11:21 +09:00
06312ad805
fix(foundry): preserve Sonnet output limits
2026-06-11 11:11:20 +09:00
c4493943e1
fix(foundry): clear API key auth on Entra setup
2026-06-11 11:11:20 +09:00
15806717d9
fix(foundry): infer selected Claude routes
2026-06-11 11:11:20 +09:00
0a8dfc21be
fix(foundry): align Claude onboarding contracts
2026-06-11 11:11:20 +09:00
b66f0b6ca3
fix(foundry): use bearer auth for Claude
2026-06-11 11:11:20 +09:00
2fa9a5eaa0
fix(foundry): scope Entra tokens by API
2026-06-11 11:11:20 +09:00
3ba617681f
fix(providers): preserve adaptive Claude routing
2026-06-11 11:11:20 +09:00
1240de7588
fix(microsoft-foundry): filter unsupported Anthropic deployments
...
Co-authored-by: Otto Deng <ottodeng@users.noreply.github.com >
2026-06-09 16:45:14 +09:00
7128fa8832
fix(image-generation): preserve explicit model defaults
2026-06-09 15:34:29 +09:00
e103d1231d
fix(microsoft-foundry): classify manual MAI image setup
2026-06-09 15:34:29 +09:00
090c492759
fix(microsoft-foundry): configure MAI image setup defaults
2026-06-09 15:34:28 +09:00
1ba782f286
feat(microsoft-foundry): classify MAI model metadata
2026-06-09 15:34:28 +09:00
248dfb22ec
fix: preserve Foundry Responses reasoning replay ids
...
Preserve Microsoft Foundry encrypted reasoning replay item ids for Responses continuations while leaving chat-completions streams untouched.
Fixes #91033 .
2026-06-07 01:24:07 -07:00
a1f1895b1b
fix(config): allow thinkingLevelMap in persisted model schema
...
Allow persisted provider model entries to carry strict thinkingLevelMap values so Microsoft Foundry Entra onboarding can save generated reasoning model config. Closes #91011 .
2026-06-06 20:44:15 -07:00
4fa5092cdc
docs: document small extension sources
2026-06-04 21:02:07 -04:00
5b0c4c0491
fix: align Foundry chat reasoning metadata
2026-06-01 02:05:38 +01:00
f13a615036
fix(foundry): bound entra token expiry clocks
2026-05-30 10:29:26 -04:00
ec43acb432
fix(microsoft-foundry): DeepSeek V4 models incorrectly use openai-completions API ( #85549 )
...
When onboarding Microsoft Foundry-hosted DeepSeek-V4 models (Pro/Flash),
the onboarding wizard assigned api: 'openai-completions' because
usesFoundryResponsesByDefault() only matched GPT/o-series models.
These V4 models require the Responses API (openai-responses) to work
correctly against the Foundry endpoint. Without this fix, all calls fail
with 'provider rejected the request schema or tool payload'.
Fix: Add 'deepseek-v4' prefix to usesFoundryResponsesByDefault() so only
the verified V4 family defaults to openai-responses. Older DeepSeek
families (e.g., V3) remain on openai-completions until proven compatible.
Closes: DeepSeek V4 models deployed via Microsoft Foundry onboarding
failing immediately due to wrong API adapter.
Co-authored-by: Roslin <rmj010203@gmail.com >
2026-05-23 15:49:53 +01:00
bc9e601491
fix: allow provider timeout overlays ( #83990 )
...
* fix: allow provider timeout overlays
* test: fix provider overlay fixture types
2026-05-21 17:10:32 -07:00
ff871e162a
fix(config): allow bundled provider timeout overlays ( #83267 )
...
* fix config provider timeout overlays
Allow bundled model provider config entries to act as overlays so fields like timeoutSeconds can be configured without redeclaring baseUrl and models. Keep unknown custom provider declarations strict, and guard configured-provider fallback against overlay entries without models.
* fix(config): include provider aliases in model overlays
* fix(config): guard Foundry timeout overlays
* fix(config): normalize bundled provider overlays
* fix(models): reject overlay-only fallback models
2026-05-18 21:50:10 -07:00
c96181fdbe
fix(foundry): wrap malformed az token json
2026-05-14 19:10:08 +08:00
5592132f56
test: dedupe foundry provider mock read
2026-05-13 05:31:55 +01:00
dea633e7dd
test: guard foundry provider mock call
2026-05-12 03:19:53 +01:00
337f1b2a42
fix(microsoft-foundry): replace unsafe non-null assertion in subscription lookup ( #62742 )
...
subs.find() can return undefined if the selected subscription ID does not match any enabled subscription. Replace the unsafe non-null assertion with an explicit guard and descriptive error.
Co-authored-by: oliviareid-svg <oliviareid@users.noreply.github.com >
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com >
2026-05-11 03:59:34 -05:00
a662afe195
test: tighten foundry provider assertions
2026-05-11 01:31:30 +01:00
827b0de0ce
refactor: reduce plugin sdk surface
2026-05-10 12:37:10 +01:00
c33d71c6b8
test: avoid extension filter allocation assertions
2026-05-08 21:35:49 +01:00
13dacceed4
test: tighten extension helper assertions
2026-05-08 20:39:25 +01:00
16a2773d4e
test: tighten microsoft foundry auth assertions
2026-05-08 19:53:49 +01:00
1ae4db279c
test: clarify foundry refresh rejection assertion
2026-05-08 07:49:58 +01:00
62ccd8b644
Fix model and tool normalization regressions
...
Summary:
- Fix model and tool normalization regressions, including explicit tool-policy grants for messaging profile warnings.
- Keep Codex and Microsoft Foundry auth handling compatible with aws-sdk auth profile modes after rebasing onto current main.
Verification:
- pnpm test src/agents/pi-tools.policy.test.ts
- pnpm tsgo:extensions
- pnpm tsgo:extensions:test
- pnpm test extensions/codex/src/app-server/auth-bridge.test.ts extensions/microsoft-foundry/index.test.ts
- pnpm test:extensions:package-boundary
- pnpm lint --threads=8
- git diff --check
- GitHub PR checks green on 4ad136106b
2026-05-07 02:29:28 -05:00
90b6665ded
refactor: move plugin api test helper to sdk
2026-04-28 00:24:54 +01:00
4336a7f3a9
refactor(plugin-sdk): narrow config runtime imports
2026-04-27 14:58:32 +01:00
d28b02a7b1
fix: preserve Foundry image capability through runtime
2026-04-05 08:36:11 -05:00
llagy007