diff --git a/CHANGELOG.md b/CHANGELOG.md index 5cf4f09dea7..80846c9d049 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -59,6 +59,7 @@ Docs: https://docs.openclaw.ai - Browser/CDP: explain that loopback Browserless or other externally managed CDP services need `attachOnly: true` and matching Browserless `EXTERNAL` endpoint when reporting local port ownership conflicts, and fall back to the configured bare WebSocket root when a discovered Browserless endpoint rejects CDP. Fixes #49815. - ACP/OpenCode: update the bundled acpx runtime to 0.6.0 and cover the OpenCode ACP bind path in Docker live tests. - Providers/OpenCode Go: add DeepSeek V4 Pro and DeepSeek V4 Flash to the Go catalog while the bundled Pi registry catches up. Fixes #71587. +- Providers/OpenCode Go: route DeepSeek V4 Pro/Flash through the OpenAI-compatible Go endpoint and suppress invalid `reasoning_effort: "off"` payloads, fixing tool-enabled requests for `opencode-go/deepseek-v4-flash`. Fixes #71683. - Browser/existing-session: support per-profile Chrome MCP command/args, map `cdpUrl` to `--browserUrl` or `--wsEndpoint`, and avoid combining endpoint flags with `--userDataDir`. Fixes #47879, #48037, and #62706. Thanks @puneet1409, @zhehao, and @madkow1001. - Media/plugins: bound MIME sniffing and ZIP archive preflight before handing untrusted files to `file-type` or `jszip`, reducing parser CPU and memory diff --git a/docs/.generated/plugin-sdk-api-baseline.sha256 b/docs/.generated/plugin-sdk-api-baseline.sha256 index 877d530c333..d414d56324f 100644 --- a/docs/.generated/plugin-sdk-api-baseline.sha256 +++ b/docs/.generated/plugin-sdk-api-baseline.sha256 @@ -1,2 +1,2 @@ -fae367b052828a57feab3bfcd10a58ebeacd6c858b337d0aab72726863952946 plugin-sdk-api-baseline.json -1bb8995e1486f7d900928aaace87421a8297fe41264197f9bf849f07c65c8f2b plugin-sdk-api-baseline.jsonl +f813474b1623f06e1465daacd56db970e8e92ab1be122faee0fa2a1dc2d4fc43 plugin-sdk-api-baseline.json +b3ea88c0c9b4cf6d9a46f0d34149063303853e78ef9708224608e4da79b23190 plugin-sdk-api-baseline.jsonl diff --git a/docs/plugins/sdk-migration.md b/docs/plugins/sdk-migration.md index 98ff9f7b100..bbf6cca7268 100644 --- a/docs/plugins/sdk-migration.md +++ b/docs/plugins/sdk-migration.md @@ -342,7 +342,7 @@ releases. | `plugin-sdk/provider-web-search` | Provider web-search helpers | Web-search provider registration/cache/runtime helpers | | `plugin-sdk/provider-tools` | Provider tool/schema compat helpers | `ProviderToolCompatFamily`, `buildProviderToolCompatFamilyHooks`, Gemini schema cleanup + diagnostics, and xAI compat helpers such as `resolveXaiModelCompatPatch` / `applyXaiModelCompat` | | `plugin-sdk/provider-usage` | Provider usage helpers | `fetchClaudeUsage`, `fetchGeminiUsage`, `fetchGithubCopilotUsage`, and other provider usage helpers | - | `plugin-sdk/provider-stream` | Provider stream wrapper helpers | `ProviderStreamFamily`, `buildProviderStreamFamilyHooks`, `composeProviderStreamWrappers`, stream wrapper types, and shared Anthropic/Bedrock/Google/Kilocode/Moonshot/OpenAI/OpenRouter/Z.A.I/MiniMax/Copilot wrapper helpers | + | `plugin-sdk/provider-stream` | Provider stream wrapper helpers | `ProviderStreamFamily`, `buildProviderStreamFamilyHooks`, `composeProviderStreamWrappers`, stream wrapper types, and shared Anthropic/Bedrock/DeepSeek V4/Google/Kilocode/Moonshot/OpenAI/OpenRouter/Z.A.I/MiniMax/Copilot wrapper helpers | | `plugin-sdk/provider-transport-runtime` | Provider transport helpers | Native provider transport helpers such as guarded fetch, transport message transforms, and writable transport event streams | | `plugin-sdk/keyed-async-queue` | Ordered async queue | `KeyedAsyncQueue` | | `plugin-sdk/media-runtime` | Shared media helpers | Media fetch/transform/store helpers plus media payload builders | diff --git a/docs/plugins/sdk-provider-plugins.md b/docs/plugins/sdk-provider-plugins.md index b34e9611b88..574fb5e5dff 100644 --- a/docs/plugins/sdk-provider-plugins.md +++ b/docs/plugins/sdk-provider-plugins.md @@ -340,7 +340,7 @@ API key auth, and dynamic model resolution. Each family builder is composed from lower-level public helpers exported from the same package, which you can reach for when a provider needs to go off the common pattern: - `openclaw/plugin-sdk/provider-model-shared` — `ProviderReplayFamily`, `buildProviderReplayFamilyHooks(...)`, and the raw replay builders (`buildOpenAICompatibleReplayPolicy`, `buildAnthropicReplayPolicyForModel`, `buildGoogleGeminiReplayPolicy`, `buildHybridAnthropicOrOpenAIReplayPolicy`). Also exports Gemini replay helpers (`sanitizeGoogleGeminiReplayHistory`, `resolveTaggedReasoningOutputMode`) and endpoint/model helpers (`resolveProviderEndpoint`, `normalizeProviderId`, `normalizeGooglePreviewModelId`, `normalizeNativeXaiModelId`). - - `openclaw/plugin-sdk/provider-stream` — `ProviderStreamFamily`, `buildProviderStreamFamilyHooks(...)`, `composeProviderStreamWrappers(...)`, plus the shared OpenAI/Codex wrappers (`createOpenAIAttributionHeadersWrapper`, `createOpenAIFastModeWrapper`, `createOpenAIServiceTierWrapper`, `createOpenAIResponsesContextManagementWrapper`, `createCodexNativeWebSearchWrapper`) and shared proxy/provider wrappers (`createOpenRouterWrapper`, `createToolStreamWrapper`, `createMinimaxFastModeWrapper`). + - `openclaw/plugin-sdk/provider-stream` — `ProviderStreamFamily`, `buildProviderStreamFamilyHooks(...)`, `composeProviderStreamWrappers(...)`, plus the shared OpenAI/Codex wrappers (`createOpenAIAttributionHeadersWrapper`, `createOpenAIFastModeWrapper`, `createOpenAIServiceTierWrapper`, `createOpenAIResponsesContextManagementWrapper`, `createCodexNativeWebSearchWrapper`), DeepSeek V4 OpenAI-compatible wrapper (`createDeepSeekV4OpenAICompatibleThinkingWrapper`), and shared proxy/provider wrappers (`createOpenRouterWrapper`, `createToolStreamWrapper`, `createMinimaxFastModeWrapper`). - `openclaw/plugin-sdk/provider-tools` — `ProviderToolCompatFamily`, `buildProviderToolCompatFamilyHooks("gemini")`, underlying Gemini schema helpers (`normalizeGeminiToolSchemas`, `inspectGeminiToolSchemas`), and xAI compat helpers (`resolveXaiModelCompatPatch()`, `applyXaiModelCompat(model)`). The bundled xAI plugin uses `normalizeResolvedModel` + `contributeResolvedModelCompat` with these to keep xAI rules owned by the provider. Some stream helpers stay provider-local on purpose. `@openclaw/anthropic-provider` keeps `wrapAnthropicProviderStream`, `resolveAnthropicBetas`, `resolveAnthropicFastMode`, `resolveAnthropicServiceTier`, and the lower-level Anthropic wrapper builders in its own public `api.ts` / `contract-api.ts` seam because they encode Claude OAuth beta handling and `context1m` gating. The xAI plugin similarly keeps native xAI Responses shaping in its own `wrapStreamFn` (`/fast` aliases, default `tool_stream`, unsupported strict-tool cleanup, xAI-specific reasoning-payload removal). diff --git a/docs/plugins/sdk-subpaths.md b/docs/plugins/sdk-subpaths.md index 8b770a6d1f8..1d5a0762a21 100644 --- a/docs/plugins/sdk-subpaths.md +++ b/docs/plugins/sdk-subpaths.md @@ -102,7 +102,7 @@ For the plugin authoring guide, see [Plugin SDK overview](/plugins/sdk-overview) | `plugin-sdk/provider-web-search` | Web-search provider registration/cache/runtime helpers | | `plugin-sdk/provider-tools` | `ProviderToolCompatFamily`, `buildProviderToolCompatFamilyHooks`, Gemini schema cleanup + diagnostics, and xAI compat helpers such as `resolveXaiModelCompatPatch` / `applyXaiModelCompat` | | `plugin-sdk/provider-usage` | `fetchClaudeUsage` and similar | - | `plugin-sdk/provider-stream` | `ProviderStreamFamily`, `buildProviderStreamFamilyHooks`, `composeProviderStreamWrappers`, stream wrapper types, and shared Anthropic/Bedrock/Google/Kilocode/Moonshot/OpenAI/OpenRouter/Z.A.I/MiniMax/Copilot wrapper helpers | + | `plugin-sdk/provider-stream` | `ProviderStreamFamily`, `buildProviderStreamFamilyHooks`, `composeProviderStreamWrappers`, stream wrapper types, and shared Anthropic/Bedrock/DeepSeek V4/Google/Kilocode/Moonshot/OpenAI/OpenRouter/Z.A.I/MiniMax/Copilot wrapper helpers | | `plugin-sdk/provider-transport-runtime` | Native provider transport helpers such as guarded fetch, transport message transforms, and writable transport event streams | | `plugin-sdk/provider-onboard` | Onboarding config patch helpers | | `plugin-sdk/global-singleton` | Process-local singleton/map/cache helpers | diff --git a/extensions/deepseek/stream.ts b/extensions/deepseek/stream.ts index 26916503bd9..4ad5bcfb544 100644 --- a/extensions/deepseek/stream.ts +++ b/extensions/deepseek/stream.ts @@ -1,76 +1,17 @@ import type { ProviderWrapStreamFnContext } from "openclaw/plugin-sdk/plugin-entry"; -import { streamWithPayloadPatch } from "openclaw/plugin-sdk/provider-stream-shared"; - -type DeepSeekThinkingLevel = ProviderWrapStreamFnContext["thinkingLevel"]; +import { createDeepSeekV4OpenAICompatibleThinkingWrapper } from "openclaw/plugin-sdk/provider-stream-shared"; function isDeepSeekV4ModelId(modelId: unknown): boolean { return modelId === "deepseek-v4-flash" || modelId === "deepseek-v4-pro"; } -function isDisabledThinkingLevel(thinkingLevel: DeepSeekThinkingLevel): boolean { - const normalized = typeof thinkingLevel === "string" ? thinkingLevel.toLowerCase() : ""; - return normalized === "off" || normalized === "none"; -} - -function resolveDeepSeekReasoningEffort(thinkingLevel: DeepSeekThinkingLevel): "high" | "max" { - return thinkingLevel === "xhigh" || thinkingLevel === "max" ? "max" : "high"; -} - -function stripDeepSeekReasoningContent(payload: Record): void { - if (!Array.isArray(payload.messages)) { - return; - } - for (const message of payload.messages) { - if (!message || typeof message !== "object") { - continue; - } - delete (message as Record).reasoning_content; - } -} - -function ensureDeepSeekToolCallReasoningContent(payload: Record): void { - if (!Array.isArray(payload.messages)) { - return; - } - for (const message of payload.messages) { - if (!message || typeof message !== "object") { - continue; - } - const record = message as Record; - if (record.role !== "assistant" || !Array.isArray(record.tool_calls)) { - continue; - } - if (!("reasoning_content" in record)) { - record.reasoning_content = ""; - } - } -} - export function createDeepSeekV4ThinkingWrapper( baseStreamFn: ProviderWrapStreamFnContext["streamFn"], - thinkingLevel: DeepSeekThinkingLevel, + thinkingLevel: ProviderWrapStreamFnContext["thinkingLevel"], ): ProviderWrapStreamFnContext["streamFn"] { - if (!baseStreamFn) { - return undefined; - } - const underlying = baseStreamFn; - return (model, context, options) => { - if (model.provider !== "deepseek" || !isDeepSeekV4ModelId(model.id)) { - return underlying(model, context, options); - } - - return streamWithPayloadPatch(underlying, model, context, options, (payload) => { - if (isDisabledThinkingLevel(thinkingLevel)) { - payload.thinking = { type: "disabled" }; - delete payload.reasoning_effort; - delete payload.reasoning; - stripDeepSeekReasoningContent(payload); - return; - } - - payload.thinking = { type: "enabled" }; - payload.reasoning_effort = resolveDeepSeekReasoningEffort(thinkingLevel); - ensureDeepSeekToolCallReasoningContent(payload); - }); - }; + return createDeepSeekV4OpenAICompatibleThinkingWrapper({ + baseStreamFn, + thinkingLevel, + shouldPatchModel: (model) => model.provider === "deepseek" && isDeepSeekV4ModelId(model.id), + }); } diff --git a/extensions/opencode-go/index.test.ts b/extensions/opencode-go/index.test.ts index eaea5ac9128..31d541e6bf6 100644 --- a/extensions/opencode-go/index.test.ts +++ b/extensions/opencode-go/index.test.ts @@ -119,15 +119,56 @@ describe("opencode-go provider plugin", () => { } as never), ).toMatchObject({ id: "deepseek-v4-pro", - api: "anthropic-messages", + api: "openai-completions", provider: "opencode-go", - baseUrl: "https://opencode.ai/zen/go", + baseUrl: "https://opencode.ai/zen/go/v1", reasoning: true, contextWindow: 1_000_000, maxTokens: 384_000, + compat: { + supportsUsageInStreaming: true, + supportsReasoningEffort: true, + maxTokensField: "max_tokens", + }, }); }); + it("disables invalid DeepSeek V4 reasoning_effort off payloads on OpenCode Go", async () => { + const provider = await registerSingleProviderPlugin(plugin); + const capturedPayloads: Record[] = []; + const baseStreamFn = (_model: unknown, _context: unknown, options: unknown) => { + const payload = { + model: "deepseek-v4-flash", + reasoning_effort: "off", + reasoning: "off", + }; + (options as { onPayload?: (payload: Record) => void })?.onPayload?.(payload); + capturedPayloads.push(payload); + return {} as never; + }; + + const streamFn = provider.wrapStreamFn?.({ + streamFn: baseStreamFn as never, + providerId: "opencode-go", + modelId: "deepseek-v4-flash", + thinkingLevel: "off", + } as never); + + expect(streamFn).toBeTypeOf("function"); + await streamFn?.( + { provider: "opencode-go", id: "deepseek-v4-flash" } as never, + {} as never, + {}, + ); + + expect(capturedPayloads).toEqual([ + { + model: "deepseek-v4-flash", + thinking: { type: "disabled" }, + }, + ]); + }); + it("canonicalizes stale OpenCode Go base URLs", async () => { const provider = await registerSingleProviderPlugin(plugin); diff --git a/extensions/opencode-go/index.ts b/extensions/opencode-go/index.ts index 926d013c8b2..6a2af044ec2 100644 --- a/extensions/opencode-go/index.ts +++ b/extensions/opencode-go/index.ts @@ -8,6 +8,7 @@ import { normalizeOpencodeGoBaseUrl, resolveOpencodeGoSupplementalModel, } from "./provider-catalog.js"; +import { createOpencodeGoDeepSeekV4Wrapper } from "./stream.js"; const PROVIDER_ID = "opencode-go"; export default definePluginEntry({ @@ -67,6 +68,7 @@ export default definePluginEntry({ resolveDynamicModel: ({ modelId }) => resolveOpencodeGoSupplementalModel(modelId), augmentModelCatalog: () => listOpencodeGoSupplementalModelCatalogEntries(), ...PASSTHROUGH_GEMINI_REPLAY_HOOKS, + wrapStreamFn: (ctx) => createOpencodeGoDeepSeekV4Wrapper(ctx.streamFn, ctx.thinkingLevel), isModernModelRef: () => true, }); api.registerMediaUnderstandingProvider(opencodeGoMediaUnderstandingProvider); diff --git a/extensions/opencode-go/provider-catalog.ts b/extensions/opencode-go/provider-catalog.ts index dce7c7d9c31..c429236dbc3 100644 --- a/extensions/opencode-go/provider-catalog.ts +++ b/extensions/opencode-go/provider-catalog.ts @@ -12,9 +12,9 @@ const OPENCODE_GO_SUPPLEMENTAL_MODELS = ( { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro", - api: "anthropic-messages", + api: "openai-completions", provider: PROVIDER_ID, - baseUrl: OPENCODE_GO_ANTHROPIC_BASE_URL, + baseUrl: OPENCODE_GO_OPENAI_BASE_URL, reasoning: true, input: ["text"], cost: { @@ -25,13 +25,18 @@ const OPENCODE_GO_SUPPLEMENTAL_MODELS = ( }, contextWindow: 1_000_000, maxTokens: 384_000, + compat: { + supportsUsageInStreaming: true, + supportsReasoningEffort: true, + maxTokensField: "max_tokens", + }, }, { id: "deepseek-v4-flash", name: "DeepSeek V4 Flash", - api: "anthropic-messages", + api: "openai-completions", provider: PROVIDER_ID, - baseUrl: OPENCODE_GO_ANTHROPIC_BASE_URL, + baseUrl: OPENCODE_GO_OPENAI_BASE_URL, reasoning: true, input: ["text"], cost: { @@ -42,6 +47,11 @@ const OPENCODE_GO_SUPPLEMENTAL_MODELS = ( }, contextWindow: 1_000_000, maxTokens: 384_000, + compat: { + supportsUsageInStreaming: true, + supportsReasoningEffort: true, + maxTokensField: "max_tokens", + }, }, ] satisfies ProviderRuntimeModel[] ).map((model) => normalizeModelCompat(model)); diff --git a/extensions/opencode-go/stream.ts b/extensions/opencode-go/stream.ts new file mode 100644 index 00000000000..b4df754786c --- /dev/null +++ b/extensions/opencode-go/stream.ts @@ -0,0 +1,18 @@ +import type { ProviderWrapStreamFnContext } from "openclaw/plugin-sdk/plugin-entry"; +import { createDeepSeekV4OpenAICompatibleThinkingWrapper } from "openclaw/plugin-sdk/provider-stream-shared"; + +function isOpencodeGoDeepSeekV4ModelId(modelId: unknown): boolean { + return modelId === "deepseek-v4-flash" || modelId === "deepseek-v4-pro"; +} + +export function createOpencodeGoDeepSeekV4Wrapper( + baseStreamFn: ProviderWrapStreamFnContext["streamFn"], + thinkingLevel: ProviderWrapStreamFnContext["thinkingLevel"], +): ProviderWrapStreamFnContext["streamFn"] { + return createDeepSeekV4OpenAICompatibleThinkingWrapper({ + baseStreamFn, + thinkingLevel, + shouldPatchModel: (model) => + model.provider === "opencode-go" && isOpencodeGoDeepSeekV4ModelId(model.id), + }); +} diff --git a/src/plugin-sdk/provider-stream-shared.ts b/src/plugin-sdk/provider-stream-shared.ts index 34135faceb2..76776a04700 100644 --- a/src/plugin-sdk/provider-stream-shared.ts +++ b/src/plugin-sdk/provider-stream-shared.ts @@ -141,6 +141,77 @@ export function createPayloadPatchStreamWrapper( ); } +export type DeepSeekV4ThinkingLevel = ProviderWrapStreamFnContext["thinkingLevel"]; + +function isDisabledDeepSeekV4ThinkingLevel(thinkingLevel: DeepSeekV4ThinkingLevel): boolean { + const normalized = typeof thinkingLevel === "string" ? thinkingLevel.toLowerCase() : ""; + return normalized === "off" || normalized === "none"; +} + +function resolveDeepSeekV4ReasoningEffort(thinkingLevel: DeepSeekV4ThinkingLevel): "high" | "max" { + return thinkingLevel === "xhigh" || thinkingLevel === "max" ? "max" : "high"; +} + +function stripDeepSeekV4ReasoningContent(payload: Record): void { + if (!Array.isArray(payload.messages)) { + return; + } + for (const message of payload.messages) { + if (!message || typeof message !== "object") { + continue; + } + delete (message as Record).reasoning_content; + } +} + +function ensureDeepSeekV4ToolCallReasoningContent(payload: Record): void { + if (!Array.isArray(payload.messages)) { + return; + } + for (const message of payload.messages) { + if (!message || typeof message !== "object") { + continue; + } + const record = message as Record; + if (record.role !== "assistant" || !Array.isArray(record.tool_calls)) { + continue; + } + if (!("reasoning_content" in record)) { + record.reasoning_content = ""; + } + } +} + +export function createDeepSeekV4OpenAICompatibleThinkingWrapper(params: { + baseStreamFn: StreamFn | undefined; + thinkingLevel: DeepSeekV4ThinkingLevel; + shouldPatchModel: (model: Parameters[0]) => boolean; +}): StreamFn | undefined { + if (!params.baseStreamFn) { + return undefined; + } + const underlying = params.baseStreamFn; + return (model, context, options) => { + if (!params.shouldPatchModel(model)) { + return underlying(model, context, options); + } + + return streamWithPayloadPatch(underlying, model, context, options, (payload) => { + if (isDisabledDeepSeekV4ThinkingLevel(params.thinkingLevel)) { + payload.thinking = { type: "disabled" }; + delete payload.reasoning_effort; + delete payload.reasoning; + stripDeepSeekV4ReasoningContent(payload); + return; + } + + payload.thinking = { type: "enabled" }; + payload.reasoning_effort = resolveDeepSeekV4ReasoningEffort(params.thinkingLevel); + ensureDeepSeekV4ToolCallReasoningContent(payload); + }); + }; +} + export type GoogleThinkingLevel = "MINIMAL" | "LOW" | "MEDIUM" | "HIGH"; export type GoogleThinkingInputLevel = | "off"