Merge branch 'main' into issue/515

Author: edelauna

.changeset/itchy-moles-thank.md

@@ -0,0 +1,5 @@
+---
+"zoo-code": patch
+---
+
+Fix bedrock DNS resolution when behind corporate proxy

.github/actions/setup-node-pnpm/action.yml

@@ -6,7 +6,7 @@ inputs:
     node-version:
         description: "Node.js version to use"
         required: false
-        default: "20.20.2"
+        default: "22.23.1"
     pnpm-version:
         description: "pnpm version to use"
         required: false

.nvmrc

@@ -1 +1 @@
-20.20.2
+22.23.1

.tool-versions

@@ -1,2 +1,2 @@
 pnpm 10.8.1
-nodejs 20.20.2
+nodejs 22.23.1

CHANGELOG.md

@@ -1,5 +1,32 @@
 # Zoo Code Changelog
 
+## [3.72.0]
+
+### Minor Changes
+
+- Add the Moonshot provider with live model discovery, streaming, model metadata, and a model picker (PR #857 by @grizmin)
+- Add the Kimi Code provider with OAuth device-flow authentication (PR #945 by @taltas)
+- Add Claude Opus 5 support across all providers (PR #1010 by @app/zoomote)
+- Add Kimi K3 to the Moonshot and OpenCode Go providers (#932 by @navedmerchant, PR #996 by @app/zoomote)
+- Add Gemini 3.6 Flash model support (PR #975 by @app/zoomote)
+- Add MiniMax-M3 model support (#888 by @RayWinter0816, PR #946 by @app/zoomote)
+- Add a safe way to abandon interrupted subtasks by severing stale parent-child links and surfacing delegation status (#559 by @edelauna, PR #935 by @edelauna)
+- Add Dart support to codebase indexing (#940 by @WebMad, PR #941 by @WebMad)
+- Fix codebase indexing for plain-text files (#931 by @tool-buddy, PR #938 by @WebMad)
+- Enable image input for DeepSeek V4 models (#964 by @grizmin, PR #963 by @grizmin)
+- Fix ChatGPT OAuth requests for GPT-5.6 Luna being rejected by the Codex backend (PR #889 by @taltas)
+- Preserve `reasoning_content` for known reasoning model families when using LiteLLM (#891 by @daewoongoh, PR #899 by @daewoongoh)
+- Fix task-history cache invalidation races by routing `invalidate()` and `invalidateAll()` through the task-history lock (#698 by @edelauna, PR #912 by @morgan-coded)
+- Fix Settings mode changes by synchronizing the local `cachedState` editing buffer (#914 by @easonLiangWorldedtech, PR #925 by @easonLiangWorldedtech)
+- Dismiss the welcome screen after successful Zoo Gateway sign-in (#961 by @JohnCanty, PR #962 by @JamesRobert20)
+- Add `CompletePromptOptions` to the `completePrompt` API so callers can configure prompt completion (#615 by @edelauna, PR #901 by @easonLiangWorldedtech)
+- Centralize provider identifiers into canonical shared types (#951 by @WebMad, PR #952 by @WebMad)
+- Refactor provider categories to use canonical provider identifiers (PR #989 by @WebMad)
+- Remove obsolete MCP server-creation translations (#895 by @edelauna, PR #943 by @WebMad)
+- Add regression coverage for resuming interrupted subtasks (#566 by @myk1yt, PR #911 by @edelauna)
+- Upload coverage reports as GitHub Actions artifacts to simplify CI debugging (PR #939 by @app/zoomote)
+- Update `esbuild-wasm` to v0.28.1 (PR #829 by @app/renovate)
+
 ## [3.70.0]
 
 ### Minor Changes

README.md

@@ -35,15 +35,15 @@
 You can find a quick guide for migrating from Roo Code to Zoo Code in the [Roo→Zoo migration guide](https://docs.zoocode.dev/roo-to-zoo-migration). We plan to try and help users as they transition over, we have our [Reddit](https://www.reddit.com/r/ZooCode) and [Discord](https://discord.gg/VxfP4Vx3gX)
 for this exact support, so if you are having problems or if you have question, jump on and ask.
 
-## What's New in v3.70.0
+## What's New in v3.72.0
 
-- **OpenAI GPT-5.6 family** — `Sol`, `Terra`, and `Luna` are now available across both the OpenAI Codex and OpenAI Native provider paths.
-- **Grok 4.5 support** — xAI's new flagship model, plus a reasoning-effort format fix that also benefits Grok 4 Mini.
-- **Kenari provider support** — a first-class, OpenAI-compatible AI gateway billed in Rupiah covering Claude, GPT, DeepSeek, GLM, Kimi and more.
-- Surface the context-compaction button and context window progress bar in the collapsed task header.
-- Fix: terminal output loss and premature task completion on cold terminals.
-- Fix: image attach for Zoo Gateway and Vercel AI Gateway models now follows live vision-capability tags instead of a static allowlist.
-- Dependency and tooling updates.
+- **Moonshot and Kimi Code providers** — discover Moonshot models dynamically or sign in to Kimi Code through its OAuth device flow.
+- **Latest model support** — use Claude Opus 5 across providers, plus Kimi K3, Gemini 3.6 Flash, and MiniMax-M3.
+- **Improved subtask workflows** — abandon interrupted subtasks cleanly, with safer task-history invalidation and better resume coverage.
+- **Expanded codebase indexing** — index Dart and plain-text files.
+- **Provider reliability** — fixes for GPT-5.6 Luna with ChatGPT OAuth, LiteLLM reasoning content, and DeepSeek V4 image input.
+- **Smoother setup and settings** — improved Zoo Gateway sign-in and mode-switch behavior.
+- API, provider-type, dependency, localization, and CI improvements.
 
 <details>
   <summary>🌐 Available languages</summary>
@@ -66,8 +66,7 @@ for this exact support, so if you are having problems or if you have question, j
 - [Tiếng Việt](locales/vi/README.md)
 - [简体中文](locales/zh-CN/README.md)
 - [繁體中文](locales/zh-TW/README.md)
-- ...
-    </details>
+</details>
 
 ---
 

apps/cli/package.json

@@ -41,7 +41,7 @@
 		"@roo-code/config-eslint": "workspace:^",
 		"@trpc/server": "^11.18.0",
 		"@roo-code/config-typescript": "workspace:^",
-		"@types/node": "20.19.43",
+		"@types/node": "22.20.1",
 		"@types/react": "18.3.31",
 		"@vitest/coverage-v8": "4.1.9",
 		"ink-testing-library": "4.0.0",

apps/vscode-e2e/package.json

@@ -17,7 +17,7 @@
 		"@roo-code/types": "workspace:^",
 		"@copilotkit/aimock": "1.35.0",
 		"@types/mocha": "10.0.10",
-		"@types/node": "20.19.43",
+		"@types/node": "22.20.1",
 		"@types/vscode": "1.100.0",
 		"@vscode/test-electron": "2.5.2",
 		"dotenv-cli": "11.0.0",

apps/vscode-e2e/src/fixtures/subtasks.ts

@@ -18,7 +18,8 @@ const SUBTASK_XPROFILE_DIFFERENT_CHILD_MARKER = "SUBTASK_CHILD_DIFFERENT_PROFILE
 const SUBTASK_CHILD_PROMPT = `${SUBTASK_CHILD_MARKER}: Ask the user exactly this follow-up question: What is the square root of 81? After the user answers, complete with only the answer.`
 export const SUBTASK_PARENT_PROMPT = `${SUBTASK_PARENT_MARKER}: Use the new_task tool exactly once. Create an ask-mode subtask with this exact message: "${SUBTASK_CHILD_PROMPT}" Do not answer directly.`
 export const SUBTASK_CHILD_FOLLOWUP_ANSWER = "9"
-const SUBTASK_FAST_CHILD_PROMPT = `${SUBTASK_FAST_CHILD_MARKER}: Complete immediately with the exact result "Fast child completed".`
+export const SUBTASK_FAST_CHILD_RESULT = "Fast child completed"
+const SUBTASK_FAST_CHILD_PROMPT = `${SUBTASK_FAST_CHILD_MARKER}: Complete immediately with the exact result "${SUBTASK_FAST_CHILD_RESULT}".`
 export const SUBTASK_FAST_PARENT_PROMPT = `${SUBTASK_FAST_PARENT_MARKER}: Use the new_task tool exactly once. Create an ask-mode subtask with this exact message: "${SUBTASK_FAST_CHILD_PROMPT}" Do not answer directly.`
 
 const SUBTASK_INTERRUPT_CHILD_PROMPT = `${SUBTASK_INTERRUPT_CHILD_MARKER}: Ask the user exactly this follow-up question: What is the square root of 81? After the user answers, complete with only the answer.`
@@ -54,6 +55,31 @@ const requestContains = (req: ChatCompletionRequest, expected: string[]) => {
 	return expected.every((text) => rawRequest.includes(text))
 }
 
+// aimock's `userMessage` matcher only inspects the LAST user message and joins only the
+// `type: "text"` content parts (see getTextContent in aimock's router). Fixtures that need
+// whole-request exclusions must replicate that scoping inside a predicate so they keep the
+// same matching semantics as the bare-regex fixtures they replace.
+const lastUserMessageContains = (req: ChatCompletionRequest, text: string) => {
+	const userMessages = req.messages?.filter((message) => message.role === "user") ?? []
+	const last = userMessages.at(-1)
+	if (!last) return false
+	const content =
+		typeof last.content === "string"
+			? last.content
+			: (last.content ?? [])
+					.filter((part): part is { type: "text"; text: string } => part?.type === "text")
+					.map((part) => part.text)
+					.join("")
+	return content.includes(text)
+}
+
+// reopenParentFromDelegation injects the child result into the resumed parent's history as
+// `Subtask <childId> completed.\n\nResult:\n<summary>`. Matching on this injected prefix (in
+// its JSON-serialized form) keeps parent-resume fixtures robust when
+// validateAndFixToolResultIds rewrites tool-use ids on resume — matching on the new_task
+// tool-call id directly proved flaky (the id can be rewritten, the fixture then misses, and
+// a looser child fixture wins and serves the child's response to the parent).
+const SUBTASK_RESULT_INJECTION = "completed.\\n\\nResult:"
 const completionAfterAnswer = (followupId: string, completionId: string) => ({
 	match: {
 		predicate: (req: ChatCompletionRequest) =>
@@ -100,25 +126,33 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 		},
 	})
 
+	// The parent prompt embeds SUBTASK_FAST_CHILD_MARKER verbatim, so parent-resume turns
+	// can also match a bare substring check (same collision class as #561). Exclude the
+	// parent marker so those turns fall through to the parent-resume fixture below.
 	mock.addFixture({
 		match: {
-			userMessage: new RegExp(SUBTASK_FAST_CHILD_MARKER),
+			predicate: (req: ChatCompletionRequest) =>
+				lastUserMessageContains(req, SUBTASK_FAST_CHILD_MARKER) &&
+				!requestContains(req, [SUBTASK_FAST_PARENT_MARKER]),
 		},
 		response: {
 			toolCalls: [
 				{
 					name: "attempt_completion",
-					arguments: JSON.stringify({ result: "Fast child completed" }),
+					arguments: JSON.stringify({ result: SUBTASK_FAST_CHILD_RESULT }),
 					id: "call_subtasks_fast_child_completion_002",
 				},
 			],
 		},
 	})
 
+	// Guard on SUBTASK_RESULT_INJECTION (not the child result text): the child result is
+	// embedded verbatim in SUBTASK_FAST_PARENT_PROMPT, so it cannot distinguish the parent's
+	// initial request from its resume turn.
 	mock.addFixture({
 		match: {
 			predicate: (req: ChatCompletionRequest) =>
-				requestContains(req, [SUBTASK_FAST_PARENT_MARKER, "call_subtasks_fast_parent_new_task_001"]),
+				requestContains(req, [SUBTASK_FAST_PARENT_MARKER, SUBTASK_RESULT_INJECTION]),
 		},
 		response: {
 			toolCalls: [
@@ -131,9 +165,15 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 		},
 	})
 
+	// This fixture is shared by several tests, so sequenceIndex cannot guard it. Exclude
+	// resume turns via the injected tool_result prefix instead: when the tool_result is
+	// serialized as role:"tool", the original parent prompt is the last user message again
+	// and would otherwise re-serve new_task on the parent's resume turn.
 	mock.addFixture({
 		match: {
-			userMessage: new RegExp(SUBTASK_PARENT_MARKER),
+			predicate: (req: ChatCompletionRequest) =>
+				lastUserMessageContains(req, SUBTASK_PARENT_MARKER) &&
+				!requestContains(req, [SUBTASK_RESULT_INJECTION]),
 		},
 		response: {
 			toolCalls: [
@@ -149,9 +189,12 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 		},
 	})
 
+	// Same collision guard as the fast-child fixture above: SUBTASK_PARENT_PROMPT embeds
+	// SUBTASK_CHILD_MARKER verbatim, so parent-resume turns must not match this fixture.
 	mock.addFixture({
 		match: {
-			userMessage: new RegExp(SUBTASK_CHILD_MARKER),
+			predicate: (req: ChatCompletionRequest) =>
+				lastUserMessageContains(req, SUBTASK_CHILD_MARKER) && !requestContains(req, [SUBTASK_PARENT_MARKER]),
 		},
 		response: {
 			toolCalls: [
@@ -172,7 +215,7 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 	mock.addFixture({
 		match: {
 			predicate: (req: ChatCompletionRequest) =>
-				requestContains(req, [SUBTASK_PARENT_MARKER, "call_subtasks_parent_new_task_001"]),
+				requestContains(req, [SUBTASK_PARENT_MARKER, SUBTASK_RESULT_INJECTION]),
 		},
 		response: {
 			toolCalls: [
@@ -238,14 +281,12 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 		},
 	})
 
+	// Same as the fast parent-resume fixture: the child result is embedded verbatim in
+	// SUBTASK_API_HANG_PARENT_PROMPT, so guard on the injected tool_result prefix instead.
 	mock.addFixture({
 		match: {
 			predicate: (req: ChatCompletionRequest) =>
-				requestContains(req, [
-					SUBTASK_API_HANG_PARENT_MARKER,
-					"call_api_hang_parent_new_task_001",
-					SUBTASK_API_HANG_CHILD_RESULT,
-				]),
+				requestContains(req, [SUBTASK_API_HANG_PARENT_MARKER, SUBTASK_RESULT_INJECTION]),
 		},
 		response: {
 			toolCalls: [
@@ -433,7 +474,7 @@ export function addSubtaskFixtures(mock: InstanceType<typeof LLMock>) {
 	mock.addFixture({
 		match: {
 			predicate: (req: ChatCompletionRequest) =>
-				requestContains(req, [SUBTASK_INTERRUPT_PARENT_MARKER, "call_interrupt_parent_new_task_001"]),
+				requestContains(req, [SUBTASK_INTERRUPT_PARENT_MARKER, SUBTASK_RESULT_INJECTION]),
 		},
 		response: {
 			toolCalls: [

apps/vscode-e2e/src/suite/subtasks.test.ts

@@ -14,6 +14,7 @@ import {
 	SUBTASK_API_HANG_PARENT_RESULT,
 	SUBTASK_API_HANG_RESUME_MESSAGE,
 	SUBTASK_CHILD_FOLLOWUP_ANSWER,
+	SUBTASK_FAST_CHILD_RESULT,
 	SUBTASK_FAST_PARENT_PROMPT,
 	SUBTASK_INTERRUPT_CHILD_FOLLOWUP_ANSWER,
 	SUBTASK_INTERRUPT_PARENT_PROMPT,
@@ -101,7 +102,8 @@ suite("Roo Code Subtasks", function () {
 					([taskId, messages]) =>
 						taskId !== parentTaskId &&
 						messages.some(
-							({ say, text }) => say === "completion_result" && text?.trim() === "Fast child completed",
+							({ say, text }) =>
+								say === "completion_result" && text?.trim() === SUBTASK_FAST_CHILD_RESULT,
 						),
 				),
 				"Immediately-completing child should emit its expected result",
@@ -357,15 +359,23 @@ suite("Roo Code Subtasks", function () {
 
 			await api.cancelCurrentTask()
 
+			// Gate on the async settle before asserting the parent never resumed: a spurious
+			// resume would be an async downstream effect of the cancellation, so a synchronous
+			// check right after cancelCurrentTask() proves nothing.
+			await waitFor(() => api.getCurrentTaskStack().at(-1) === spawnedTaskId)
+			await waitFor(
+				() => asks[spawnedTaskId!]?.some(({ type, ask }) => type === "ask" && ask === "resume_task") ?? false,
+			)
+
 			assert.ok(
 				messages[parentTaskId]?.find(({ type, text }) => type === "say" && text === "Parent task resumed") ===
 					undefined,
 				"Parent task should not have resumed after subtask cancellation",
 			)
-
-			await waitFor(() => api.getCurrentTaskStack().at(-1) === spawnedTaskId)
-			await waitFor(
-				() => asks[spawnedTaskId!]?.some(({ type, ask }) => type === "ask" && ask === "resume_task") ?? false,
+			assert.strictEqual(
+				messages[parentTaskId]?.find(({ say }) => say === "completion_result"),
+				undefined,
+				"Parent must not have completed after subtask cancellation",
 			)
 
 			await api.clearCurrentTask()
@@ -539,15 +549,15 @@ suite("Roo Code Subtasks", function () {
 				says[childTaskId!]
 					?.filter(({ say }) => say === "completion_result")
 					.map(({ text }) => text?.trim())
-					.find((text) => text === SUBTASK_API_HANG_CHILD_RESULT),
+					.find((text): text is string => !!text),
 				SUBTASK_API_HANG_CHILD_RESULT,
 				"Child should complete with its expected result after resume",
 			)
 			assert.strictEqual(
 				says[parentTaskId]
 					?.filter(({ say }) => say === "completion_result")
 					.map(({ text }) => text?.trim())
-					.find((text) => text === SUBTASK_API_HANG_PARENT_RESULT),
+					.find((text): text is string => !!text),
 				SUBTASK_API_HANG_PARENT_RESULT,
 				"Parent should resume and complete with its expected result",
 			)