@cloudcli-ai/cloudcli 1.37.1 → 1.37.3

package/server/modules/providers/list/codex/codex.provider.ts

@@ -1,5 +1,6 @@
 import { AbstractProvider } from '@/modules/providers/shared/base/abstract.provider.js';
 import { CodexProviderAuth } from '@/modules/providers/list/codex/codex-auth.provider.js';
+import { CodexForkProvider } from '@/modules/providers/list/codex/codex-fork.provider.js';
 import { CodexProviderModels } from '@/modules/providers/list/codex/codex-models.provider.js';
 import { codexRuntime } from '@/modules/providers/list/codex/codex-runtime.provider.js';
 import { CodexMcpProvider } from '@/modules/providers/list/codex/codex-mcp.provider.js';
@@ -8,6 +9,7 @@ import { CodexSessionsProvider } from '@/modules/providers/list/codex/codex-sess
 import { CodexSkillsProvider } from '@/modules/providers/list/codex/codex-skills.provider.js';
 import type {
   IProviderAuth,
+  IProviderFork,
   IProviderModels,
   IProviderRuntime,
   IProviderSessionSynchronizer,
@@ -22,6 +24,7 @@ export class CodexProvider extends AbstractProvider {
   readonly auth: IProviderAuth = new CodexProviderAuth();
   readonly skills: IProviderSkills = new CodexSkillsProvider();
   readonly sessions: IProviderSessions = new CodexSessionsProvider();
+  readonly fork: IProviderFork = new CodexForkProvider();
   readonly sessionSynchronizer: IProviderSessionSynchronizer = new CodexSessionSynchronizer();
 
   constructor() {

package/server/modules/providers/list/cursor/cursor-runtime.provider.js

@@ -30,22 +30,24 @@ function isWorkspaceTrustPrompt(text = '') {
 }
 
 async function spawnCursor(command, options = {}, ws, context) {
-  return new Promise(async (resolve, reject) => {
+  // Callers pass the stable app session id; the CLI resumes with the
+  // provider-native id recorded on the session row. Both lookups run before the
+  // promise is created: inside an async executor a rejected `resolveResumeModel`
+  // would be swallowed and leave the returned promise pending forever.
+  const providerSessionId = context.resolveProviderSessionId(options.sessionId);
+  const resolvedModel = await context.resolveResumeModel(options.sessionId, options.model);
+
+  return new Promise((resolve, reject) => {
     const {
       sessionId,
       projectPath,
       cwd,
       toolsSettings,
       skipPermissions,
-      model,
       sessionSummary,
       images,
       files
     } = options;
-    // Callers pass the stable app session id; the CLI resumes with the
-    // provider-native id recorded on the session row.
-    const providerSessionId = context.resolveProviderSessionId(sessionId);
-    const resolvedModel = await context.resolveResumeModel(sessionId, model);
     let capturedSessionId = providerSessionId; // Track the provider-native session id throughout the process
     let sessionCreatedSent = false; // Track if we've already sent session-created event
     let hasRetriedWithTrust = false;

package/server/modules/providers/list/cursor/cursor-session-synchronizer.provider.ts

@@ -1,4 +1,3 @@
-import crypto from 'node:crypto';
 import fs from 'node:fs';
 import fsp from 'node:fs/promises';
 import os from 'node:os';

package/server/modules/providers/list/opencode/opencode-models.provider.ts

@@ -1,3 +1,7 @@
+import { readFile } from 'node:fs/promises';
+import os from 'node:os';
+import path from 'node:path';
+
 import Database from 'better-sqlite3';
 
 import { sessionsDb } from '@/modules/database/index.js';
@@ -17,9 +21,9 @@ import {
  * Curated OpenCode catalog shipped as immutable CloudCLI defaults.
  *
  * OpenCode routes by `<providerID>/<modelID>`, so this list mirrors the
- * providers `opencode models --verbose` reports: the OpenCode Zen gateway plus
- * the Anthropic and OpenAI providers OpenCode can address directly with the
- * user's own credentials.
+ * providers `opencode models --verbose` reports: the OpenCode Zen gateway, the
+ * OpenCode Go subscription gateway, and the Anthropic and OpenAI providers
+ * OpenCode can address directly with the user's own credentials.
  */
 export const OPENCODE_PREDEFINED_MODELS: ProviderModelsDefinition = {
   OPTIONS: [
@@ -77,6 +81,178 @@ export const OPENCODE_PREDEFINED_MODELS: ProviderModelsDefinition = {
     { value: 'opencode/north-mini-code-free', label: 'North Mini Code Free', description: 'OpenCode Zen · Free' },
     { value: 'opencode/nemotron-3-ultra-free', label: 'Nemotron 3 Ultra Free', description: 'OpenCode Zen · Free' },
     { value: 'opencode/deepseek-v4-flash-free', label: 'DeepSeek V4 Flash Free', description: 'OpenCode Zen · Free' },
+    {
+      value: 'opencode-go/grok-4.6',
+      label: 'Grok 4.6',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'medium' }, { value: 'high' }, { value: 'xhigh' }],
+      },
+    },
+    {
+      value: 'opencode-go/glm-5.3-flash',
+      label: 'GLM 5.3 Flash',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'high' }, { value: 'max' }],
+      },
+    },
+    {
+      value: 'opencode-go/glm-5.3',
+      label: 'GLM 5.3',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'high' }, { value: 'max' }],
+      },
+    },
+    {
+      value: 'opencode-go/glm-5.2',
+      label: 'GLM 5.2',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'high' }, { value: 'max' }],
+      },
+    },
+    { value: 'opencode-go/glm-5.1', label: 'GLM 5.1', description: 'OpenCode Go' },
+    {
+      value: 'opencode-go/gpt-5.6-luna',
+      label: 'GPT 5.6 Luna',
+      description: 'OpenCode Go',
+      effort: {
+        values: [
+          { value: 'none' },
+          { value: 'low' },
+          { value: 'medium' },
+          { value: 'high' },
+          { value: 'xhigh' },
+          { value: 'max' },
+        ],
+      },
+    },
+    {
+      value: 'opencode-go/kimi-k3',
+      label: 'Kimi K3',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'max' }],
+      },
+    },
+    { value: 'opencode-go/kimi-k2.7-code', label: 'Kimi K2.7 Code', description: 'OpenCode Go' },
+    { value: 'opencode-go/kimi-k2.6', label: 'Kimi K2.6', description: 'OpenCode Go' },
+    {
+      value: 'opencode-go/longcat-2.0',
+      label: 'LongCat 2.0',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'medium' }, { value: 'high' }],
+      },
+    },
+    { value: 'opencode-go/mimo-v2.5', label: 'MiMo V2.5', description: 'OpenCode Go' },
+    { value: 'opencode-go/mimo-v2.5-pro', label: 'MiMo V2.5 Pro', description: 'OpenCode Go' },
+    {
+      value: 'opencode-go/minimax-m3',
+      label: 'MiniMax M3',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'none' }, { value: 'thinking' }],
+      },
+    },
+    { value: 'opencode-go/minimax-m2.7', label: 'MiniMax M2.7', description: 'OpenCode Go' },
+    {
+      value: 'opencode-go/muse-spark-1.3-contributor',
+      label: 'Muse Spark 1.3 Contributor',
+      description: 'OpenCode Go',
+      effort: {
+        values: [
+          { value: 'minimal' },
+          { value: 'low' },
+          { value: 'medium' },
+          { value: 'high' },
+          { value: 'xhigh' },
+        ],
+      },
+    },
+    {
+      value: 'opencode-go/muse-spark-1.2-contributor',
+      label: 'Muse Spark 1.2 Contributor',
+      description: 'OpenCode Go',
+      effort: {
+        values: [
+          { value: 'minimal' },
+          { value: 'low' },
+          { value: 'medium' },
+          { value: 'high' },
+          { value: 'xhigh' },
+        ],
+      },
+    },
+    {
+      value: 'opencode-go/qwen3.8-max',
+      label: 'Qwen3.8 Max',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'medium' }, { value: 'xhigh' }],
+      },
+    },
+    {
+      value: 'opencode-go/qwen3.8-flash',
+      label: 'Qwen3.8 Flash',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'medium' }, { value: 'xhigh' }],
+      },
+    },
+    { value: 'opencode-go/qwen3.7-max', label: 'Qwen3.7 Max', description: 'OpenCode Go' },
+    { value: 'opencode-go/qwen3.7-plus', label: 'Qwen3.7 Plus', description: 'OpenCode Go' },
+    { value: 'opencode-go/qwen3.6-plus', label: 'Qwen3.6 Plus', description: 'OpenCode Go' },
+    {
+      value: 'opencode-go/deepseek-v4-pro',
+      label: 'DeepSeek V4 Pro',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'high' }, { value: 'max' }],
+      },
+    },
+    {
+      value: 'opencode-go/deepseek-v4-flash',
+      label: 'DeepSeek V4 Flash',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'high' }, { value: 'max' }],
+      },
+    },
+    {
+      value: 'opencode-go/deepseek-v4-flash-vision-exp',
+      label: 'DeepSeek V4 Flash Vision Exp',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'high' }, { value: 'max' }],
+      },
+    },
+    {
+      value: 'opencode-go/hy4-preview',
+      label: 'Hy4 Preview',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'none' }, { value: 'high' }],
+      },
+    },
+    {
+      value: 'opencode-go/hy3',
+      label: 'Hy3',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'none' }, { value: 'low' }, { value: 'high' }],
+      },
+    },
+    {
+      value: 'opencode-go/omen-alpha',
+      label: 'Omen Alpha',
+      description: 'OpenCode Go',
+      effort: {
+        values: [{ value: 'low' }, { value: 'high' }],
+      },
+    },
     { value: 'anthropic/claude-opus-5', label: 'Claude Opus 5', description: 'Anthropic' },
     { value: 'anthropic/claude-opus-5-fast', label: 'Claude Opus 5 Fast', description: 'Anthropic' },
     { value: 'anthropic/claude-fable-5', label: 'Claude Fable 5', description: 'Anthropic' },
@@ -117,6 +293,100 @@ export const OPENCODE_PREDEFINED_MODELS: ProviderModelsDefinition = {
   DEFAULT: 'opencode/gpt-5.6-terra',
 };
 
+/** Global OpenCode config files, in the order the CLI loads them. */
+const OPENCODE_CONFIG_FILES = ['config.json', 'opencode.json', 'opencode.jsonc'];
+
+/** Provider API keys OpenCode reads straight from the environment. */
+const OPENCODE_ENV_PROVIDER_IDS: Record<string, string> = {
+  OPENCODE_API_KEY: 'opencode',
+  ANTHROPIC_API_KEY: 'anthropic',
+  OPENAI_API_KEY: 'openai',
+};
+
+const readOpenCodeJsonFile = async (filePath: string): Promise<Record<string, unknown> | null> => {
+  try {
+    return readObjectRecord(JSON.parse(await readFile(filePath, 'utf8')));
+  } catch {
+    // Missing, unreadable, or comment-bearing (.jsonc) files simply contribute
+    // nothing; the auth store is the authoritative source below.
+    return null;
+  }
+};
+
+/**
+ * Lists the upstream providers this OpenCode install can actually route to.
+ *
+ * OpenCode resolves `<providerID>/<modelID>` against the providers the user has
+ * connected, and rejects anything else outright - `Model
+ * opencode/claude-sonnet-4-6 is not valid` is what a run gets for asking for an
+ * OpenCode Zen model on a machine that only has an Anthropic key. The curated
+ * catalog spans every provider OpenCode can address, so it has to be narrowed
+ * to this machine's providers before it reaches the model picker.
+ *
+ * Returns null when nothing can be read, so the caller keeps the full catalog
+ * rather than leaving the picker empty. Providers declared only in a
+ * project-level `opencode.json` are not visible here; the null fallback and the
+ * env-key sweep keep those installs on the full list.
+ */
+const readConnectedOpenCodeProviderIds = async (): Promise<Set<string> | null> => {
+  const providerIds = new Set<string>();
+  const configDir = path.join(os.homedir(), '.config', 'opencode');
+
+  const auth = await readOpenCodeJsonFile(
+    path.join(os.homedir(), '.local', 'share', 'opencode', 'auth.json'),
+  );
+  for (const [providerId, credential] of Object.entries(auth ?? {})) {
+    if (readObjectRecord(credential)) {
+      providerIds.add(providerId);
+    }
+  }
+
+  for (const configFile of OPENCODE_CONFIG_FILES) {
+    const config = await readOpenCodeJsonFile(path.join(configDir, configFile));
+    for (const providerId of Object.keys(readObjectRecord(config?.provider) ?? {})) {
+      providerIds.add(providerId);
+    }
+  }
+
+  for (const [envKey, providerId] of Object.entries(OPENCODE_ENV_PROVIDER_IDS)) {
+    if (readOptionalString(process.env[envKey])) {
+      providerIds.add(providerId);
+    }
+  }
+
+  return providerIds.size > 0 ? providerIds : null;
+};
+
+/**
+ * Narrows the curated catalog to the providers OpenCode can route to.
+ *
+ * The default has to move with the list: leaving it on an OpenCode Zen model
+ * would hand every new session a model the CLI refuses to run.
+ */
+const filterOpenCodeModelsByProvider = (
+  definition: ProviderModelsDefinition,
+  connectedProviderIds: Set<string> | null,
+): ProviderModelsDefinition => {
+  if (!connectedProviderIds) {
+    return definition;
+  }
+
+  const options = definition.OPTIONS.filter(
+    (option) => connectedProviderIds.has(option.value.split('/')[0]),
+  );
+  if (options.length === 0) {
+    return definition;
+  }
+
+  return {
+    ...definition,
+    OPTIONS: options,
+    DEFAULT: options.some((option) => option.value === definition.DEFAULT)
+      ? definition.DEFAULT
+      : options[0].value,
+  };
+};
+
 const parseOpenCodeSessionModelValue = (rawModel: unknown): string | null => {
   if (typeof rawModel === 'string') {
     const trimmed = rawModel.trim();
@@ -146,12 +416,15 @@ const parseOpenCodeSessionModelValue = (rawModel: unknown): string | null => {
 /** Provider registry model adapter for OpenCode predefined models and session metadata. */
 export class OpenCodeProviderModels implements IProviderModels {
   async getSupportedModels(): Promise<ProviderModelsDefinition> {
-    return OPENCODE_PREDEFINED_MODELS;
+    return filterOpenCodeModelsByProvider(
+      OPENCODE_PREDEFINED_MODELS,
+      await readConnectedOpenCodeProviderIds(),
+    );
   }
 
   async getCurrentActiveModel(sessionId?: string): Promise<ProviderCurrentActiveModel> {
     if (!sessionId?.trim()) {
-      return buildDefaultProviderCurrentActiveModel(OPENCODE_PREDEFINED_MODELS);
+      return buildDefaultProviderCurrentActiveModel(await this.getSupportedModels());
     }
 
     // OpenCode's `session` table is keyed by its own session id, so the stable
@@ -198,6 +471,6 @@ export class OpenCodeProviderModels implements IProviderModels {
       // Fall through to the curated default when OpenCode session lookup fails.
     }
 
-    return buildDefaultProviderCurrentActiveModel(OPENCODE_PREDEFINED_MODELS);
+    return buildDefaultProviderCurrentActiveModel(await this.getSupportedModels());
   }
 }

package/server/modules/providers/provider.routes.ts

@@ -297,6 +297,28 @@ const parseProvider = (value: unknown): LLMProvider => {
   });
 };
 
+/** Both fields are optional: an empty body forks the whole conversation. */
+const parseSessionForkPayload = (payload: unknown): { upToAnchorId?: string; title?: string } => {
+  if (payload === undefined || payload === null) {
+    return {};
+  }
+  if (typeof payload !== 'object') {
+    throw new AppError('Request body must be an object.', {
+      code: 'INVALID_REQUEST_BODY',
+      statusCode: 400,
+    });
+  }
+
+  const body = payload as Record<string, unknown>;
+  const upToAnchorId = typeof body.upToAnchorId === 'string' ? body.upToAnchorId.trim() : '';
+  const title = typeof body.title === 'string' ? body.title.trim() : '';
+
+  return {
+    ...(upToAnchorId ? { upToAnchorId } : {}),
+    ...(title ? { title } : {}),
+  };
+};
+
 const parseSessionRenameSummary = (payload: unknown): string => {
   if (!payload || typeof payload !== 'object') {
     throw new AppError('Request body must be an object.', {
@@ -792,6 +814,15 @@ router.post(
   }),
 );
 
+router.post(
+  '/sessions/:sessionId/fork',
+  asyncHandler(async (req: Request, res: Response) => {
+    const sessionId = parseSessionId(req.params.sessionId);
+    const result = await sessionsService.forkSessionById(sessionId, parseSessionForkPayload(req.body));
+    res.status(201).json(createApiSuccessResponse(result));
+  }),
+);
+
 router.put(
   '/sessions/:sessionId',
   asyncHandler(async (req: Request, res: Response) => {
@@ -840,6 +871,11 @@ router.get('/search/sessions', asyncHandler(async (req: Request, res: Response)
       query,
       limit,
       signal: abortController.signal,
+      onTitleResults: (titleResults) => {
+        if (!closed) {
+          res.write(`event: title-results\ndata: ${JSON.stringify({ titleResults })}\n\n`);
+        }
+      },
       onProgress: ({ projectResult, totalMatches, scannedProjects, totalProjects }) => {
         if (closed) {
           return;

package/server/modules/providers/services/provider-capabilities.service.ts

@@ -25,6 +25,15 @@ type ProviderCapabilities = {
   supportsTokenUsage: boolean;
   /** Whether the provider runtime can accept model-level reasoning effort. */
   supportsEffort: boolean;
+  /**
+   * Whether an already-sent message can be replaced, which requires the
+   * provider to re-run a conversation truncated at a chosen point.
+   */
+  supportsMessageEditing: boolean;
+  /**
+   * Whether a session's transcript can be branched into an independent one.
+   */
+  supportsSessionForking: boolean;
 };
 
 /**
@@ -44,6 +53,10 @@ const PROVIDER_CAPABILITIES: Record<LLMProvider, ProviderCapabilities> = {
     supportsPermissionRequests: true,
     supportsTokenUsage: true,
     supportsEffort: true,
+    // `resumeSessionAt` re-runs a conversation truncated at a message, and
+    // `forkSession` copies a transcript prefix into a new session file.
+    supportsMessageEditing: true,
+    supportsSessionForking: true,
   },
   cursor: {
     provider: 'cursor',
@@ -55,6 +68,8 @@ const PROVIDER_CAPABILITIES: Record<LLMProvider, ProviderCapabilities> = {
     supportsPermissionRequests: false,
     supportsTokenUsage: false,
     supportsEffort: false,
+    supportsMessageEditing: false,
+    supportsSessionForking: false,
   },
   codex: {
     provider: 'codex',
@@ -66,6 +81,12 @@ const PROVIDER_CAPABILITIES: Record<LLMProvider, ProviderCapabilities> = {
     supportsPermissionRequests: false,
     supportsTokenUsage: true,
     supportsEffort: true,
+    // Not from the Codex SDK, which only starts and resumes threads: both ride
+    // the same CLI's `app-server` protocol, whose `thread/fork` copies a
+    // thread up to a chosen turn. Editing is that fork plus a new prompt,
+    // which is how Codex's own IDE clients do it.
+    supportsMessageEditing: true,
+    supportsSessionForking: true,
   },
   opencode: {
     provider: 'opencode',
@@ -80,6 +101,8 @@ const PROVIDER_CAPABILITIES: Record<LLMProvider, ProviderCapabilities> = {
     supportsPermissionRequests: false,
     supportsTokenUsage: true,
     supportsEffort: true,
+    supportsMessageEditing: false,
+    supportsSessionForking: false,
   },
 };