Feature/pinned chats

[andrii-ivanov]

PR Type

Enhancement, Tests

---

Description

• Pinned Conversations Feature: Added comprehensive support for pinning and reordering conversations with dedicated UI components and state management
• MCP (Model Context Protocol) Enhancements: Significantly improved MCP server management with OAuth authentication, connection handling, and comprehensive test coverage
• Theme System Implementation: Added complete theming infrastructure with dark/light themes, RGB color definitions, and theme application utilities
• Test Coverage Expansion: Added extensive test suites for MCP functionality, access middleware, tool formatting, memory agents, and OpenAI configurations
• Configuration Schema Updates: Enhanced endpoint configurations with new parameters like verbosity, web_search, SearXNG support, and file search capabilities
• Code Quality Improvements: Refactored various hooks and utilities for better state management, removed deprecated components, and improved error handling
• Internationalization Setup: Added comprehensive i18n configuration supporting 26 languages with proper fallback strategies
• Memory Agent Enhancements: Improved memory functionality with overflow handling and GPT-5+ model support

---

Diagram Walkthrough

flowchart LR
  A["Pinned Conversations"] --> B["Conversation Grouping"]
  A --> C["Reorder Mutations"]
  D["MCP Enhancements"] --> E["OAuth Handler"]
  D --> F["Server Manager"]
  D --> G["Connection Status"]
  H["Theme System"] --> I["Dark Theme"]
  H --> J["Light Theme"]
  H --> K["Theme Utils"]
  L["Test Coverage"] --> M["MCP Tests"]
  L --> N["Access Tests"]
  L --> O["Tool Tests"]
  P["Configuration"] --> Q["SearXNG Support"]
  P --> R["Web Search"]
  P --> S["File Search"]

Loading

File Walkthrough

	Relevant files
Tests	13 files zod.spec.ts Enhanced MCP Zod schema tests with ref resolution                packages/api/src/mcp/zod.spec.ts • Updated imports to use convertWithResolvedRefs instead of convertJsonSchemaToZod throughout test file • Added comprehensive tests for $ref resolution functionality with resolveJsonSchemaRefs • Added tests for handling bare object schemas as passthrough for dynamic properties • Added tests for additionalProperties with anyOf/oneOf and allowEmptyObject scenarios +776/-44 web.spec.ts Updated web search tests with new imports and SearXNG support packages/api/src/web/web.spec.ts • Updated import paths to use relative imports and librechat-data-provider • Added searxngInstanceUrl and searxngApiKey fields to all test configurations • Fixed TypeScript type annotations for authFields.forEach callback parameters • Added comprehensive tests for firecrawlOptions properties and timeout handling +330/-23 access.spec.ts Added comprehensive access middleware test suite                  packages/api/src/middleware/access.spec.ts • Added comprehensive test suite for access control middleware functions • Tests checkAccess, generateCheckAccess, and skipAgentCheck functions • Covers permission validation, role-based access, and real-world usage patterns • Includes error handling and edge case scenarios for access control +554/-0  format.spec.ts Added comprehensive tool formatting test suite                      packages/api/src/tools/format.spec.ts • Added comprehensive test suite for tool formatting utility functions • Tests filterUniquePlugins, checkPluginAuth, convertMCPToolsToPlugins, and getToolkitKey • Covers plugin deduplication, authentication validation, and MCP tool conversion • Includes edge cases and error handling for tool formatting operations +461/-0  mcp.spec.ts Enhanced MCP tests with HTTP type and GitHub server support packages/api/src/mcp/mcp.spec.ts • Added tests for "http" type alias support in StreamableHTTPOptionsSchema • Added tests for processing customUserVars in args field and headers • Added GitHub MCP server configuration tests with PAT_TOKEN placeholder handling • Enhanced test coverage for HTTP-based MCP server configurations +146/-0  memory.test.ts Add comprehensive memory agent test suite                                packages/api/src/agents/tests/memory.test.ts • Added comprehensive test suite for memory agent functionality • Tests memory tool creation with overflow handling and token limits • Tests GPT-5+ model handling in processMemory function with temperature removal • Tests various model formats and parameter transformations +469/-0  llm.spec.ts Add comprehensive OpenAI LLM configuration test suite        packages/api/src/endpoints/openai/llm.spec.ts • Added comprehensive test suite for getOpenAIConfig function • Tests model options, parameter separation, reasoning params, and Azure configuration • Tests GPT-5+ model handling with token parameter transformations • Tests verbosity parameter handling and various edge cases +424/-0  crud.spec.ts Add Mistral file deletion and cleanup tests                            packages/api/src/files/mistral/crud.spec.ts • Added deleteMistralFile function tests with error handling • Added comprehensive file cleanup tests for OCR processing • Tests cleanup behavior on success, failure, and error scenarios • Tests graceful handling of deletion errors without throwing +386/-1  content.spec.ts Add content formatting test suite                                                packages/api/src/format/content.spec.ts • Added test suite for formatContentStrings function • Tests conversion of message content arrays to strings for different message types • Tests handling of mixed content types and edge cases • Tests real-world scenarios with text and image content +340/-0  key.test.ts Add service key loading test suite                                              packages/api/src/utils/key.test.ts • Added comprehensive test suite for loadServiceKey function • Tests JSON parsing, file loading, URL fetching, and base64 decoding • Tests private key formatting and escaped newline handling • Tests error scenarios and edge cases +182/-0  handler.test.ts Add MCP OAuth handler configuration tests                                packages/api/src/mcp/oauth/handler.test.ts • Adds comprehensive test suite for MCP OAuth handler configuration • Tests custom OAuth metadata fields like grant_types_supported, token_endpoint_auth_methods_supported • Validates default values and custom configurations for OAuth parameters +190/-0  auth.test.ts Add MCP authentication mapping tests                                          packages/api/src/mcp/auth.test.ts • Tests MCP authentication mapping functionality with various server name formats • Validates handling of special characters, spaces, and Unicode in server names • Tests edge cases like empty tools arrays and database error handling +168/-0  convos.spec.ts Add pinned conversation grouping tests                                      client/src/utils/convos.spec.ts • Adds comprehensive tests for pinned conversation handling in groupConversationsByDate • Tests pinned conversation sorting by pinnedOrder and proper grouping • Validates edge cases with no pinned conversations +80/-0
Enhancement	33 files useMCPServerManager.ts Added comprehensive MCP server management hook                      client/src/hooks/MCP/useMCPServerManager.ts • Created comprehensive hook for managing MCP server connections and authentication • Implements OAuth flow handling with polling, cancellation, and timeout management • Provides server state management, configuration dialog handling, and batch operations • Includes connection status monitoring and user plugin authentication management +566/-0  manager.ts Refactor MCP manager initialization and OAuth handling      packages/api/src/mcp/manager.ts • Refactored initializeMCP method to extract single server initialization logic • Added tracking for OAuth servers that require authentication at startup • Modified OAuth handling to skip authentication during startup for better reliability • Added utility methods for getting user connections and OAuth servers +212/-168 connection.ts Improve MCP connection verification and error handling      packages/api/src/mcp/connection.ts • Improved connection verification logic with fallback methods • Enhanced isConnected method to handle servers without ping support • Removed error emission and cache invalidation logic • Fixed transport type detection for HTTP connections +57/-103 memory.ts Enhance memory agent with overflow handling and GPT-5+ support packages/api/src/agents/memory.ts • Enhanced memory tool with token overflow detection and error artifacts • Added GPT-5+ model parameter handling with temperature removal • Updated memory instructions to be more restrictive about when to store memories • Improved error handling and artifact generation for memory operations +92/-38  useEventHandlers.ts Improve SSE event handling and conversation state management client/src/hooks/SSE/useEventHandlers.ts • Enhanced error handling for cancelled streams and blank page scenarios • Improved conversation state management and navigation logic • Added draft saving for interrupted conversations • Fixed agent template application timing and parameters +68/-25  llm.ts Enhance OpenAI LLM configuration with GPT-5+ and tool support packages/api/src/endpoints/openai/llm.ts • Added known OpenAI parameters set for better parameter separation • Enhanced GPT-5+ model handling with token parameter transformations • Added web search tool support and verbosity parameter handling • Improved Azure configuration with Responses API support +145/-6  convos.ts Add pinned conversations support to conversation grouping client/src/utils/convos.ts • Added pinned conversations grouping and sorting functionality • Enhanced groupConversationsByDate to handle pinned conversations separately • Added pinned conversations at the top with custom ordering • Maintained existing date-based grouping for unpinned conversations +40/-6    index.ts Add comprehensive theme type definitions                                  packages/client/src/theme/types/index.ts • Defines comprehensive theme type interfaces for RGB color values, CSS variables, and theme colors • Includes interfaces for IThemeRGB, IThemeVariables, IThemeColors, and Theme • Covers text, surface, border, brand, and utility color definitions +189/-0  zod.ts Enhance JSON schema to Zod conversion with reference resolution packages/api/src/mcp/zod.ts • Removes inline type definitions and imports them from external types module • Adds resolveJsonSchemaRefs function for handling JSON schema references • Enhances schema conversion with better handling of bare object schemas and additional properties • Adds convertWithResolvedRefs helper function for tests +109/-18 object-traverse.ts Add ESM-native object traversal utility                                    packages/data-schemas/src/utils/object-traverse.ts • Implements ESM-native object traversal utility with forEach functionality • Provides comprehensive traversal context with path tracking, circular reference detection • Includes type guards and safe property manipulation methods +178/-0  mutations.ts Enhance conversation pinning and reordering mutations        client/src/data-provider/mutations.ts • Updates conversation pinning mutation to use updateConversation instead of dedicated pinConversation • Adds useReorderPinnedConversationsMutation for managing pinned conversation order • Enhances archive mutation to automatically unpin conversations when archiving +88/-26  schemas.ts Expand model parameters and conversation schema enhancements packages/data-provider/src/schemas.ts • Adds minimal option to ReasoningEffort enum and Verbosity enum • Updates penalty parameter ranges from 0-2 to -2-2 for OpenAI and agents settings • Adds web_search configuration for Anthropic settings and schema updates • Adds pinnedOrder field to conversation schema and removes inline isPinned definition +37/-12  parsers.ts Replace traverse library with custom ESM implementation    packages/data-schemas/src/config/parsers.ts • Replaces traverse library with custom ESM-native object traversal utility • Updates debug logging formatter to use new traversal context interface • Improves error handling and string concatenation in logging +54/-44  parameterSettings.ts Add new model parameters and web search options                    packages/data-provider/src/parameterSettings.ts • Adds minimal option to reasoning effort settings and verbosity parameter for OpenAI • Adds web_search toggle for OpenAI and Anthropic endpoints • Adds disableStreaming option for OpenAI endpoints • Updates parameter configurations across multiple model endpoints +74/-5    dark.ts Add dark theme color definitions                                                  packages/client/src/theme/themes/dark.ts • Defines complete dark theme color palette with RGB values • Maps colors for text, surfaces, borders, brand elements, and utility colors • Provides comprehensive dark mode color scheme implementation +72/-0    default.ts Add default light theme color definitions                                packages/client/src/theme/themes/default.ts • Defines complete default light theme color palette with RGB values • Maps colors for text, surfaces, borders, brand elements, and utility colors • Provides comprehensive light mode color scheme implementation +72/-0    applyTheme.ts Add theme application utility with validation                        packages/client/src/theme/utils/applyTheme.ts • Implements theme application utility with RGB validation • Maps theme RGB values to CSS variables and applies them to document root • Includes validation for RGB format and error handling +115/-0  handler.ts Enhance MCP OAuth handler with configurable metadata          packages/api/src/mcp/oauth/handler.ts • Updates OAuth metadata discovery to use discoverAuthorizationServerMetadata • Adds configurable OAuth metadata fields support with default values • Enhances authorization flow with resource parameter handling • Improves error handling and logging throughout OAuth process +52/-9    format.ts Add plugin formatting and conversion utilities                      packages/api/src/tools/format.ts • Adds utility functions for filtering unique plugins and checking plugin authentication • Implements MCP tools to plugins conversion functionality • Adds toolkit key resolution for tool name matching +142/-0  useToolToggle.ts Refactor tool toggle state management                                        client/src/hooks/Plugins/useToolToggle.ts • Refactors tool toggle logic to use ephemeral agent as single source of truth • Improves localStorage synchronization and removes infinite loop prevention • Enhances type safety with ToolValue type and better value handling +38/-26  useArtifacts.ts Update artifacts hook to use dedicated context                      client/src/hooks/Artifacts/useArtifacts.ts • Updates to use useArtifactsContext instead of useChatContext • Changes from latestMessage to latestMessageId and latestMessageText • Simplifies artifact detection and management logic +21/-18  crud.ts Add file cleanup and fix service key configuration              packages/api/src/files/mistral/crud.ts • Adds deleteMistralFile function for cleaning up uploaded files • Updates OCR upload process to automatically delete temporary files after processing • Fixes Google service key path environment variable name +49/-5    index.ts Consolidate SVG icon exports                                                          packages/client/src/svgs/index.ts • Consolidates all SVG icon exports into a single index file • Exports 67 different SVG components for consistent icon usage +67/-0    index.ts Simplify component SVG exports to file-related icons          client/src/components/svg/index.ts • Reduces exports to only file-related SVG components • Exports CodePaths, FileIcon, FilePaths, SheetPaths, and TextPaths +5/-66    useAutoSave.ts Refactor auto-save to use centralized draft utilities        client/src/hooks/Input/useAutoSave.ts • Refactors to use external getDraft, setDraft, and clearDraft utilities • Removes inline base64 encoding/decoding functions • Simplifies draft handling logic with centralized utilities +9/-46    i18n.ts Add comprehensive internationalization configuration          packages/client/src/locales/i18n.ts • Sets up comprehensive i18n configuration with 26 language translations • Configures language detection and fallback strategies • Includes support for Chinese variants and proper fallback chains +87/-0    key.ts Enhance Google service key loading with multiple formats  packages/api/src/utils/key.ts • Enhances Google service key loading to support base64 encoded and stringified JSON • Adds proper private key formatting with newline handling • Improves PEM format validation and correction +51/-5    useMCPSelect.ts Enhance MCP selection with configuration-based filtering  client/src/hooks/Plugins/useMCPSelect.ts • Adds filtering based on startup config chatMenu setting • Stabilizes setMCPValues callback to prevent infinite render loops • Integrates with startup configuration for MCP server visibility +23/-4    web.ts Add SearXNG search provider support                                            packages/api/src/web/web.ts • Adds SearXNG search provider configuration with instance URL and API key • Updates web search authentication to include SearXNG provider • Enhances scraper timeout configuration with Firecrawl options support +21/-5    useChatFunctions.ts Simplify chat functions by removing artifacts configuration client/src/hooks/Chat/useChatFunctions.ts • Removes artifacts mode configuration from chat functions • Simplifies message submission by removing artifact-related parameters • Updates response message ID generation logic for regeneration +4/-11    react-query-service.ts Add MCP server management mutations and queries                    packages/data-provider/src/react-query/react-query-service.ts • Adds useReinitializeMCPServerMutation for server reinitialization • Adds useCancelMCPOAuthMutation for OAuth cancellation • Adds useMCPServerConnectionStatusQuery for connection status monitoring +53/-0    files.ts Update file utility imports for SVG components                      client/src/utils/files.ts • Imports file-related SVG components from the component SVG module +1/-4      role.ts Add file search permission to role types                                  packages/data-schemas/src/types/role.ts • Adds FILE_SEARCH permission type to role interface • Extends permissions structure with file search capabilities +3/-0
Configuration changes	1 files config.ts Enhance configuration schemas with new options                      packages/data-provider/src/config.ts • Added new title-related configuration options to base endpoint schema • Added fileSearch interface configuration option • Enhanced web search schema with SearXNG and Firecrawl options • Added memory character limit configuration and updated cache keys +112/-28
Additional files	101 files .env.example +43/-3    backend-review.yml +1/-1      client.yml +58/-0    data-schemas.yml +1/-1      frontend-review.yml +1/-1      generate-release-changelog-pr.yml +0/-95    generate-unreleased-changelog-pr.yml +0/-107  i18n-unused-keys.yml +2/-1      locize-i18n-sync.yml +1/-1      unused-packages.yml +95/-6    Dockerfile +1/-1      Dockerfile.multi +11/-2    README.md +1/-1      BaseClient.js +14/-3    OpenAIClient.js +3/-1      formatMessages.js +0/-32    BaseClient.test.js +40/-0    DALLE3.js +7/-3      OpenAIImageTools.js +22/-3    DALLE3-proxy.spec.js +94/-0    fileSearch.js +10/-2    handleTools.js +15/-11  cacheConfig.js +62/-0    cacheConfig.spec.js +157/-0  cacheFactory.js +108/-0  cacheFactory.spec.js +432/-0  getLogStores.js +54/-111 ioredisClient.js +0/-92    keyvRedis.js +0/-109  logViolation.js +3/-2      redisClients.js +204/-0  Agent.js +11/-13  Agent.spec.js +40/-53  Conversation.js +19/-11  Conversation.spec.js +572/-0  File.js +113/-3  File.spec.js +264/-0  Message.js +1/-1      Message.spec.js +269/-4  tx.js +15/-2    tx.spec.js +104/-3  package.json +13/-10  ErrorController.js +12/-6    ErrorController.spec.js +241/-0  PluginController.js +26/-55  PluginController.spec.js +520/-0  TwoFactorController.js +26/-0    UserController.js +14/-16  client.js +196/-42 client.test.js +1190/-0 errors.js +0/-12    request.js +22/-0    v1.js +43/-22  v1.spec.js +681/-0  chatV1.js +3/-3      chatV2.js +1/-1      errors.js +2/-2      helpers.js +10/-0    v1.js +2/-2      tools.js +1/-2      index.js +5/-3      index.spec.js +24/-1    abortRun.js +2/-2      buildEndpointOption.js +1/-1      checkBan.js +1/-1      concurrentLimiter.js +2/-2      forkLimiters.js +79/-0    importLimiters.js +13/-22  index.js +2/-0      loginLimiter.js +5/-14    messageLimiters.js +7/-21    registerLimiter.js +5/-14    resetPasswordLimiter.js +3/-13    sttLimiters.js +7/-21    toolCallLimiter.js +5/-14    ttsLimiters.js +7/-21    uploadLimiters.js +7/-20    verifyEmailLimiter.js +3/-13    uaParser.js +3/-2      validateEndpoint.js +1/-1      validateModel.js +1/-1      mcp.spec.js +1259/-0 static.spec.js +162/-0  config.js +8/-1      convos.js +47/-5    files.agents.test.js +282/-0  files.js +129/-12 files.test.js +302/-0  mcp.js +399/-9  memories.js +74/-18  static.js +4/-1      AppService.js +5/-3      AppService.spec.js +205/-3  update.js +20/-5    update.spec.js +180/-2  AssistantService.js +6/-4      getCustomConfig.js +36/-40  getEndpointsConfig.js +12/-2    loadAsyncEndpoints.js +19/-19  loadCustomConfig.js +8/-6      Additional files not shown

---

[github-actions]

🚨 Unused NPM Packages Detected

The following unused dependencies were found:

📂 Root package.json

• @e2b/code-interpreter

⚠️ Please remove these unused dependencies to keep your project clean.

[qodo-code-review]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 4 🔵🔵🔵🔵⚪

🧪 PR contains tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Test Coverage Shift Many tests now rely on convertWithResolvedRefs and resolveJsonSchemaRefs. Ensure these helpers are robust and exported as used; verify behavior parity with the old convertJsonSchemaToZod where refs aren’t needed, and confirm unresolved-ref paths are intentionally different. describe('additionalProperties with anyOf/oneOf and allowEmptyObject', () => { it('should handle anyOf with object containing only additionalProperties when allowEmptyObject is false', () => { const schema: JsonSchemaType & { anyOf?: any } = { type: 'object', properties: { filter: { description: 'Filter field', anyOf: [ { type: 'object', additionalProperties: { type: 'object', properties: { _icontains: { type: 'string' }, }, }, }, { type: 'null', }, ], } as JsonSchemaType & { anyOf?: any }, }, }; const zodSchema = convertWithResolvedRefs(schema, { allowEmptyObject: false, transformOneOfAnyOf: true, }); expect(zodSchema).toBeDefined(); const testData = { filter: { title: { _icontains: 'Pirate', }, }, }; const result = zodSchema?.parse(testData); expect(result).toEqual(testData); expect(result?.filter).toBeDefined(); expect(result?.filter?.title?._icontains).toBe('Pirate'); }); it('should not treat objects with additionalProperties as empty', () => { const schema: JsonSchemaType = { type: 'object', additionalProperties: { type: 'string', }, }; const zodSchemaWithoutAllow = convertWithResolvedRefs(schema, { allowEmptyObject: false, }); // Should not return undefined because it has additionalProperties expect(zodSchemaWithoutAllow).toBeDefined(); const testData = { customField: 'value', }; expect(zodSchemaWithoutAllow?.parse(testData)).toEqual(testData); }); it('should handle oneOf with object containing only additionalProperties', () => { const schema: JsonSchemaType & { oneOf?: any } = { type: 'object', properties: {}, oneOf: [ { type: 'object', additionalProperties: true, }, { type: 'object', properties: { specificField: { type: 'string' }, }, }, ], }; const zodSchema = convertWithResolvedRefs(schema, { allowEmptyObject: false, transformOneOfAnyOf: true, }); expect(zodSchema).toBeDefined(); // Test with additional properties const testData1 = { randomField: 'value', anotherField: 123, }; expect(zodSchema?.parse(testData1)).toEqual(testData1); // Test with specific field const testData2 = { specificField: 'test', }; expect(zodSchema?.parse(testData2)).toEqual(testData2); }); it('should handle complex nested schema with $ref-like structure', () => { const schema: JsonSchemaType & { anyOf?: any } = { type: 'object', properties: { query: { type: 'object', properties: { filter: { description: 'Filter conditions', anyOf: [ { // This simulates a resolved $ref anyOf: [ { type: 'object', properties: { _or: { type: 'array', items: { type: 'object' }, }, }, required: ['_or'], }, { type: 'object', additionalProperties: { anyOf: [ { type: 'object', properties: { _icontains: { type: 'string' }, _eq: { type: 'string' }, }, }, ], }, }, ], }, { type: 'null', }, ], } as JsonSchemaType & { anyOf?: any }, }, }, }, }; const zodSchema = convertWithResolvedRefs(schema, { allowEmptyObject: false, transformOneOfAnyOf: true, }); expect(zodSchema).toBeDefined(); const testData = { query: { filter: { title: { _icontains: 'Pirate', }, }, }, }; const result = zodSchema?.parse(testData); expect(result).toEqual(testData); expect(result?.query?.filter?.title?._icontains).toBe('Pirate'); }); }); describe('$ref resolution with resolveJsonSchemaRefs', () => { it('should handle schemas with $ref references when resolved', () => { const schemaWithRefs = { type: 'object' as const, properties: { collection: { type: 'string' as const, }, query: { type: 'object' as const, properties: { filter: { anyOf: [{ $ref: '#/$defs/__schema0' }, { type: 'null' as const }], }, }, }, }, required: ['collection', 'query'], $defs: { __schema0: { anyOf: [ { type: 'object' as const, properties: { _or: { type: 'array' as const, items: { $ref: '#/$defs/__schema0' }, }, }, required: ['_or'], }, { type: 'object' as const, additionalProperties: { anyOf: [ { type: 'object' as const, properties: { _eq: { anyOf: [ { type: 'string' as const }, { type: 'number' as const }, { type: 'null' as const }, ], }, }, }, ], }, }, ], }, }, }; // First test without resolving refs - should not work properly // Intentionally NOT using convertWithResolvedRefs here to test the behavior without ref resolution const zodSchemaUnresolved = convertJsonSchemaToZod(schemaWithRefs as any, { allowEmptyObject: true, transformOneOfAnyOf: true, }); const testData = { collection: 'posts', query: { filter: { status: { _eq: 'draft', }, }, }, }; // Without resolving refs, the filter field won't work correctly const resultUnresolved = zodSchemaUnresolved?.parse(testData); expect(resultUnresolved?.query?.filter).toEqual({}); // Now resolve refs first const resolvedSchema = resolveJsonSchemaRefs(schemaWithRefs); // Verify refs were resolved expect(resolvedSchema.properties?.query?.properties?.filter?.anyOf?.[0]).not.toHaveProperty( '$ref', ); expect(resolvedSchema.properties?.query?.properties?.filter?.anyOf?.[0]).toHaveProperty( 'anyOf', ); // Already resolved manually above, so we use convertJsonSchemaToZod directly const zodSchemaResolved = convertJsonSchemaToZod(resolvedSchema as any, { allowEmptyObject: true, transformOneOfAnyOf: true, }); // With resolved refs, it should work correctly const resultResolved = zodSchemaResolved?.parse(testData); expect(resultResolved).toEqual(testData); expect(resultResolved?.query?.filter?.status?._eq).toBe('draft'); }); it('should handle circular $ref references without infinite loops', () => { const schemaWithCircularRefs = { type: 'object' as const, properties: { node: { $ref: '#/$defs/TreeNode' }, }, $defs: { TreeNode: { type: 'object' as const, properties: { value: { type: 'string' as const }, children: { type: 'array' as const, items: { $ref: '#/$defs/TreeNode' }, }, }, }, }, }; // Should not throw or hang const resolved = resolveJsonSchemaRefs(schemaWithCircularRefs); expect(resolved).toBeDefined(); // The circular reference should be broken with a simple object schema // Already resolved manually above, so we use convertJsonSchemaToZod directly const zodSchema = convertJsonSchemaToZod(resolved as any, { allowEmptyObject: true, transformOneOfAnyOf: true, }); expect(zodSchema).toBeDefined(); const testData = { node: { value: 'root', children: [ { value: 'child1', children: [], }, ], }, }; expect(() => zodSchema?.parse(testData)).not.toThrow(); }); it('should handle various edge cases safely', () => { // Test with null/undefined expect(resolveJsonSchemaRefs(null as any)).toBeNull(); expect(resolveJsonSchemaRefs(undefined as any)).toBeUndefined(); // Test with non-object primitives expect(resolveJsonSchemaRefs('string' as any)).toBe('string'); expect(resolveJsonSchemaRefs(42 as any)).toBe(42); expect(resolveJsonSchemaRefs(true as any)).toBe(true); // Test with arrays const arrayInput = [{ type: 'string' }, { $ref: '#/def' }]; const arrayResult = resolveJsonSchemaRefs(arrayInput as any); expect(Array.isArray(arrayResult)).toBe(true); expect(arrayResult).toHaveLength(2); // Test with schema that has no refs const noRefSchema = { type: 'object' as const, properties: { name: { type: 'string' as const }, nested: { type: 'object' as const, properties: { value: { type: 'number' as const }, }, }, }, }; const resolvedNoRef = resolveJsonSchemaRefs(noRefSchema); expect(resolvedNoRef).toEqual(noRefSchema); // Test with invalid ref (non-existent) const invalidRefSchema = { type: 'object' as const, properties: { item: { $ref: '#/$defs/nonExistent' }, }, $defs: { other: { type: 'string' as const }, }, }; const resolvedInvalid = resolveJsonSchemaRefs(invalidRefSchema); // Invalid refs should be preserved as-is expect(resolvedInvalid.properties?.item?.$ref).toBe('#/$defs/nonExistent'); // Test with empty object expect(resolveJsonSchemaRefs({})).toEqual({}); // Test with schema containing special JSON Schema keywords const schemaWithKeywords = { type: 'object' as const, properties: { value: { type: 'string' as const, minLength: 5, maxLength: 10, pattern: '^[A-Z]', }, }, additionalProperties: false, minProperties: 1, }; const resolvedKeywords = resolveJsonSchemaRefs(schemaWithKeywords); expect(resolvedKeywords).toEqual(schemaWithKeywords); expect(resolvedKeywords.properties?.value?.minLength).toBe(5); expect(resolvedKeywords.additionalProperties).toBe(false); }); }); describe('Bare object schema handling for dynamic properties', () => { it('should handle object type without explicit properties but expecting dynamic field definitions', () => { // This simulates the Kintone add_fields tool schema const schema: JsonSchemaType = { type: 'object', properties: { app_id: { type: 'number', description: 'アプリID', }, properties: { type: 'object', description: 'フィールドの設定（各フィールドには code, type, label の指定が必須）', }, }, required: ['app_id', 'properties'], }; const zodSchema = convertWithResolvedRefs(schema); // Test case 1: Basic field definition const testData1 = { app_id: 810, properties: { minutes_id: { code: 'minutes_id', type: 'SINGLE_LINE_TEXT', label: 'minutes_id', }, }, }; // WITH THE FIX: Bare object schemas now act as passthrough const result = zodSchema?.parse(testData1); expect(result).toEqual(testData1); // Properties pass through! }); it('should work when properties field has additionalProperties true', () => { const schema: JsonSchemaType = { type: 'object', properties: { app_id: { type: 'number', description: 'アプリID', }, properties: { type: 'object', description: 'フィールドの設定（各フィールドには code, type, label の指定が必須）', additionalProperties: true, }, }, required: ['app_id', 'properties'], }; const zodSchema = convertWithResolvedRefs(schema); const testData = { app_id: 810, properties: { minutes_id: { code: 'minutes_id', type: 'SINGLE_LINE_TEXT', label: 'minutes_id', }, }, }; const result = zodSchema?.parse(testData); expect(result).toEqual(testData); expect(result?.properties?.minutes_id).toBeDefined(); }); it('should work with proper field type definitions in additionalProperties', () => { const schema: JsonSchemaType = { type: 'object', properties: { app_id: { type: 'number', description: 'アプリID', }, properties: { type: 'object', description: 'フィールドの設定（各フィールドには code, type, label の指定が必須）', additionalProperties: { type: 'object', properties: { type: { type: 'string' }, code: { type: 'string' }, label: { type: 'string' }, required: { type: 'boolean' }, options: { type: 'object', additionalProperties: { type: 'object', properties: { label: { type: 'string' }, index: { type: 'string' }, }, }, }, }, required: ['type', 'code', 'label'], }, }, }, required: ['app_id', 'properties'], }; const zodSchema = convertWithResolvedRefs(schema); // Test case 1: Simple text field const testData1 = { app_id: 810, properties: { minutes_id: { code: 'minutes_id', type: 'SINGLE_LINE_TEXT', label: 'minutes_id', required: false, }, }, }; const result1 = zodSchema?.parse(testData1); expect(result1).toEqual(testData1); // Test case 2: Dropdown field with options const testData2 = { app_id: 820, properties: { status: { type: 'DROP_DOWN', code: 'status', label: 'Status', options: { 'Not Started': { label: 'Not Started', index: '0', }, 'In Progress': { label: 'In Progress', index: '1', }, }, }, }, }; const result2 = zodSchema?.parse(testData2); expect(result2).toEqual(testData2); // Test case 3: Multiple fields const testData3 = { app_id: 123, properties: { number_field: { type: 'NUMBER', code: 'number_field', label: '数値フィールド', }, text_field: { type: 'SINGLE_LINE_TEXT', code: 'text_field', label: 'テキストフィールド', }, }, }; const result3 = zodSchema?.parse(testData3); expect(result3).toEqual(testData3); }); it('should handle the actual reported failing case', () => { // This is the exact schema that's failing for the user const schema: JsonSchemaType = { type: 'object', properties: { app_id: { type: 'number', description: 'アプリID', }, properties: { type: 'object', description: 'フィールドの設定（各フィールドには code, type, label の指定が必須）', }, }, required: ['app_id', 'properties'], }; const zodSchema = convertWithResolvedRefs(schema); // The exact data the user is trying to send const userData = { app_id: 810, properties: { minutes_id: { code: 'minutes_id', type: 'SINGLE_LINE_TEXT', label: 'minutes_id', required: false, }, }, }; // WITH THE FIX: The properties now pass through correctly! const result = zodSchema?.parse(userData); expect(result).toEqual(userData); // This fixes the error "properties requires at least one field definition" // The MCP server now receives the full properties object }); it('should demonstrate fix by treating bare object type as passthrough', () => { // Test what happens if we modify the conversion to treat bare object types // without properties as passthrough schemas const schema: JsonSchemaType = { type: 'object', properties: { app_id: { type: 'number', description: 'アプリID', }, properties: { type: 'object', description: 'フィールドの設定（各フィールドには code, type, label の指定が必須）', }, }, required: ['app_id', 'properties'], }; // For now, we'll simulate the fix by adding additionalProperties const fixedSchema: JsonSchemaType = { ...schema, properties: { ...schema.properties, properties: { ...(schema.properties!.properties as JsonSchemaType), additionalProperties: true, }, }, }; const zodSchema = convertWithResolvedRefs(fixedSchema); const userData = { app_id: 810, properties: { minutes_id: { code: 'minutes_id', type: 'SINGLE_LINE_TEXT', label: 'minutes_id', required: false, }, }, }; const result = zodSchema?.parse(userData); expect(result).toEqual(userData); }); it('should NOT treat object schemas with $ref or complex properties as bare objects', () => { // This test ensures our fix doesn't affect schemas with $ref or other complex structures const schemaWithRef = { type: 'object' as const, properties: { data: { type: 'object' as const, // This has anyOf with $ref - should NOT be treated as a bare object anyOf: [{ $ref: '#/$defs/dataSchema' }, { type: 'null' as const }], }, }, $defs: { dataSchema: { type: 'object' as const, additionalProperties: { type: 'string' as const, }, }, }, }; // Convert without resolving refs const zodSchema = convertJsonSchemaToZod(schemaWithRef as any, { transformOneOfAnyOf: true, }); const testData = { data: { field1: 'value1', field2: 'value2', }, }; // Without ref resolution, the data field should be stripped/empty const result = zodSchema?.parse(testData); expect(result?.data).toEqual({}); }); it('should NOT treat object schemas with oneOf/anyOf as bare objects', () => { // Ensure schemas with oneOf/anyOf are not treated as bare objects const schemaWithOneOf = { type: 'object' as const, properties: { config: { type: 'object' as const, // Empty properties but has oneOf - should NOT be passthrough oneOf: [ { properties: { type: { const: 'A' } } }, { properties: { type: { const: 'B' } } }, ], } as any, }, }; const zodSchema = convertWithResolvedRefs(schemaWithOneOf as any, { transformOneOfAnyOf: true, }); const testData = { config: { randomField: 'should not pass through', }, }; // The random field should be stripped because this isn't a bare object const result = zodSchema?.parse(testData); expect(result?.config).toEqual({}); }); }); }); Interval Cleanup Polling uses setInterval and stores handles in state; verify all intervals are reliably cleared on success, timeout, error, and on unmount. Consider a useEffect cleanup to clear any remaining intervals to avoid leaks. const startServerPolling = useCallback( (serverName: string) => { const pollInterval = setInterval(async () => { try { await queryClient.refetchQueries([QueryKeys.mcpConnectionStatus]); const freshConnectionData = queryClient.getQueryData([ QueryKeys.mcpConnectionStatus, ]) as any; const freshConnectionStatus = freshConnectionData?.connectionStatus || {}; const state = serverStates[serverName]; const serverStatus = freshConnectionStatus[serverName]; if (serverStatus?.connectionState === 'connected') { clearInterval(pollInterval); showToast({ message: localize('com_ui_mcp_authenticated_success', { 0: serverName }), status: 'success', }); const currentValues = mcpValuesRef.current ?? []; if (!currentValues.includes(serverName)) { setMCPValues([...currentValues, serverName]); } await queryClient.invalidateQueries([QueryKeys.tools]); // This delay is to ensure UI has updated with new connection status before cleanup // Otherwise servers will show as disconnected for a second after OAuth flow completes setTimeout(() => { cleanupServerState(serverName); }, 1000); return; } if (state?.oauthStartTime && Date.now() - state.oauthStartTime > 180000) { showToast({ message: localize('com_ui_mcp_oauth_timeout', { 0: serverName }), status: 'error', }); clearInterval(pollInterval); cleanupServerState(serverName); return; } if (serverStatus?.connectionState === 'error') { showToast({ message: localize('com_ui_mcp_init_failed'), status: 'error', }); clearInterval(pollInterval); cleanupServerState(serverName); return; } } catch (error) { console.error(`[MCP Manager] Error polling server ${serverName}:`, error); clearInterval(pollInterval); cleanupServerState(serverName); return; } }, 3500); updateServerState(serverName, { pollInterval }); }, [ queryClient, serverStates, showToast, localize, setMCPValues, cleanupServerState, updateServerState, ], ); Backward Compatibility Schema changes add fields (verbosity, web_search, disableStreaming) and move isPinned/pinnedOrder. Validate API consumers and persistence layers tolerate these additions and optionality, and that min bounds for penalties changing to -2 don’t break UI validation. export enum Verbosity { none = '', low = 'low', medium = 'medium', high = 'high', } export const imageDetailNumeric = { [ImageDetail.low]: 0, [ImageDetail.auto]: 1,

[qodo-code-review]

PR Code Suggestions ✨

Explore these optional code suggestions:

Category	Suggestion	Impact
Possible issue	Reuse exact redirect URI Ensure redirectUri matches the value used during startAuthorization to prevent authorization code mismatch errors. Persist and reuse the exact redirect URI chosen for the flow instead of recomputing it here. packages/api/src/mcp/oauth/handler.ts [372-379] +const redirectUri = + metadata.redirectUri ?? + metadata.clientInfo.redirect_uris?.[0] ?? + this.getDefaultRedirectUri(); + const tokens = await exchangeAuthorization(metadata.serverUrl, { - redirectUri: metadata.clientInfo.redirect_uris?.[0] || this.getDefaultRedirectUri(), + redirectUri, metadata: metadata.metadata as unknown as SDKOAuthMetadata, clientInformation: metadata.clientInfo, codeVerifier: metadata.codeVerifier, authorizationCode, resource, }); [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 9 __ Why: This is a critical fix for the OAuth flow, as using a mismatched redirect_uri during token exchange would cause the authorization to fail.	High
	Verify connection before counting initialized Avoid marking a server as initialized solely based on a resolved function call. Verify the underlying connection state before counting it successful to prevent false positives when OAuth is required or timeouts occur. packages/api/src/mcp/manager.ts [44-84] public async initializeMCPs({ mcpServers, flowManager, tokenMethods, }: { mcpServers: t.MCPServers; flowManager: FlowStateManager<MCPOAuthTokens | null>; tokenMethods?: TokenMethods; }): Promise<void> { this.mcpConfigs = mcpServers; const entries = Object.entries(mcpServers); - const initializedServers = new Set(); + const initializedServers = new Set<number>(); const connectionResults = await Promise.allSettled( entries.map(async ([serverName, config], i) => { try { await this.initializeMCP({ serverName, config, flowManager, tokenMethods, }); - initializedServers.add(i); + const conn = this.connections.get(serverName); + if (conn && (await conn.isConnected())) { + initializedServers.add(i); + } else { + logger.warn(`[MCP][${serverName}] Skipping success count: not connected`); + } } catch (error) { logger.error(`[MCP][${serverName}] Initialization failed`, error); } }), ); - const successfulConnections = connectionResults.filter((result) => result.status === 'fulfilled').length; + const successfulConnections = initializedServers.size; logger.info(`[MCP] Initialized ${successfulConnections}/${entries.length} servers`); if (successfulConnections === 0) { throw new Error('No MCP servers initialized successfully'); } } [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 8 __ Why: The suggestion correctly identifies that initializeMCP can resolve without throwing an error even if the connection fails (e.g., due to skipped OAuth), leading to an incorrect count of successful initializations. Checking conn.isConnected() ensures only truly active connections are counted as successful, which is a critical fix for the new logic.	Medium
	Fix polling state staleness and leaks Avoid capturing stale serverStates inside the interval callback. Read the latest state via a ref or pass timeout start in closure. Also ensure the interval is cleared on unmount to prevent leaks if the component unmounts during polling. client/src/hooks/MCP/useMCPServerManager.ts [136-211] +const serverStatesRef = useRef(serverStates); +useEffect(() => { + serverStatesRef.current = serverStates; +}, [serverStates]); + +useEffect(() => { + return () => { + // cleanup all pending intervals on unmount + Object.values(serverStatesRef.current).forEach((s) => { + if (s?.pollInterval) clearInterval(s.pollInterval); + }); + }; +}, []); + const startServerPolling = useCallback( (serverName: string) => { const pollInterval = setInterval(async () => { try { await queryClient.refetchQueries([QueryKeys.mcpConnectionStatus]); - - const freshConnectionData = queryClient.getQueryData([ - QueryKeys.mcpConnectionStatus, - ]) as any; + const freshConnectionData = queryClient.getQueryData([QueryKeys.mcpConnectionStatus]) as any; const freshConnectionStatus = freshConnectionData?.connectionStatus || {}; - const state = serverStates[serverName]; + const state = serverStatesRef.current[serverName]; const serverStatus = freshConnectionStatus[serverName]; if (serverStatus?.connectionState === 'connected') { clearInterval(pollInterval); - - showToast({ - message: localize('com_ui_mcp_authenticated_success', { 0: serverName }), - status: 'success', - }); - + showToast({ message: localize('com_ui_mcp_authenticated_success', { 0: serverName }), status: 'success' }); const currentValues = mcpValuesRef.current ?? []; if (!currentValues.includes(serverName)) { setMCPValues([...currentValues, serverName]); } - await queryClient.invalidateQueries([QueryKeys.tools]); - - // This delay is to ensure UI has updated with new connection status before cleanup - // Otherwise servers will show as disconnected for a second after OAuth flow completes - setTimeout(() => { - cleanupServerState(serverName); - }, 1000); + setTimeout(() => cleanupServerState(serverName), 1000); return; } if (state?.oauthStartTime && Date.now() - state.oauthStartTime > 180000) { - showToast({ - message: localize('com_ui_mcp_oauth_timeout', { 0: serverName }), - status: 'error', - }); + showToast({ message: localize('com_ui_mcp_oauth_timeout', { 0: serverName }), status: 'error' }); clearInterval(pollInterval); cleanupServerState(serverName); return; } if (serverStatus?.connectionState === 'error') { - showToast({ - message: localize('com_ui_mcp_init_failed'), - status: 'error', - }); + showToast({ message: localize('com_ui_mcp_init_failed'), status: 'error' }); clearInterval(pollInterval); cleanupServerState(serverName); return; } } catch (error) { console.error(`[MCP Manager] Error polling server ${serverName}:`, error); clearInterval(pollInterval); cleanupServerState(serverName); return; } }, 3500); updateServerState(serverName, { pollInterval }); }, - [ - queryClient, - serverStates, - showToast, - localize, - setMCPValues, - cleanupServerState, - updateServerState, - ], + [queryClient, showToast, localize, setMCPValues, cleanupServerState, updateServerState], ); [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly identifies a stale state capture in the setInterval callback and a missing cleanup for the interval, providing a robust fix using a ref and a useEffect for unmount cleanup.	Medium
	Safely guard SharedArrayBuffer access Accessing SharedArrayBuffer can throw or be undefined in some environments. Use a typeof/safe check before referencing the constructor to avoid ReferenceErrors in browsers without support. packages/data-schemas/src/utils/object-traverse.ts [43] -if (value instanceof SharedArrayBuffer) return false; +// Check for SharedArrayBuffer when available +// eslint-disable-next-line no-undef +if (typeof SharedArrayBuffer !== 'undefined' && value instanceof SharedArrayBuffer) { + return false; +} [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 7 __ Why: The suggestion correctly prevents a potential ReferenceError in environments where SharedArrayBuffer is not available, improving the utility's robustness and portability.	Medium
	Guard against duplicate initialization Prevent duplicate initialization when multiple clicks occur. Disable re-entry per server while isInitializing or an OAuth flow is active to avoid racing mutations and multiple tabs opening. client/src/hooks/MCP/useMCPServerManager.ts [213-276] const initializeServer = useCallback( async (serverName: string, autoOpenOAuth: boolean = true) => { + const currentState = serverStatesRef.current[serverName]; + if (currentState?.isInitializing || currentState?.oauthUrl) { + return; + } + updateServerState(serverName, { isInitializing: true }); - try { const response = await reinitializeMutation.mutateAsync(serverName); if (response.success) { if (response.oauthRequired && response.oauthUrl) { updateServerState(serverName, { oauthUrl: response.oauthUrl, oauthStartTime: Date.now(), isCancellable: true, isInitializing: true, }); if (autoOpenOAuth) { window.open(response.oauthUrl, '_blank', 'noopener,noreferrer'); } startServerPolling(serverName); } else { await queryClient.refetchQueries([QueryKeys.mcpConnectionStatus]); + showToast({ message: localize('com_ui_mcp_initialized_success', { 0: serverName }), status: 'success' }); - showToast({ - message: localize('com_ui_mcp_initialized_success', { 0: serverName }), - status: 'success', - }); - - const currentValues = mcpValues ?? []; + const currentValues = mcpValuesRef.current ?? []; if (!currentValues.includes(serverName)) { setMCPValues([...currentValues, serverName]); } cleanupServerState(serverName); } } else { - showToast({ - message: localize('com_ui_mcp_init_failed', { 0: serverName }), - status: 'error', - }); + showToast({ message: localize('com_ui_mcp_init_failed', { 0: serverName }), status: 'error' }); cleanupServerState(serverName); } } catch (error) { console.error(`[MCP Manager] Failed to initialize ${serverName}:`, error); - showToast({ - message: localize('com_ui_mcp_init_failed', { 0: serverName }), - status: 'error', - }); + showToast({ message: localize('com_ui_mcp_init_failed', { 0: serverName }), status: 'error' }); cleanupServerState(serverName); } }, - [ - updateServerState, - reinitializeMutation, - startServerPolling, - queryClient, - showToast, - localize, - mcpValues, - cleanupServerState, - setMCPValues, - ], + [updateServerState, reinitializeMutation, startServerPolling, queryClient, showToast, localize, setMCPValues, cleanupServerState], ); [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 6 __ Why: The suggestion correctly identifies a potential race condition and adds a guard to prevent initializeServer from being called multiple times, which improves the hook's robustness against rapid user input.	Low
	Sanitize tool value types Guard against non-boolean/string values stored in ephemeralAgent[toolKey] to avoid unexpected UI states and localStorage pollution. Coerce only booleans and strings, and ignore objects/numbers by falling back to false. client/src/hooks/Plugins/useToolToggle.ts [75-85] -const toolValue = useMemo(() => { - return ephemeralAgent?.[toolKey] ?? false; +const toolValue = useMemo<ToolValue | false>(() => { + const value = ephemeralAgent?.[toolKey]; + if (typeof value === 'boolean' || typeof value === 'string') { + return value as ToolValue; + } + return false; }, [ephemeralAgent, toolKey]); const isToolEnabled = useMemo(() => { - // For backward compatibility, treat truthy string values as enabled if (typeof toolValue === 'string') { return toolValue.length > 0; } return toolValue === true; }, [toolValue]); [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 6 __ Why: The suggestion improves robustness by ensuring the toolValue is always a boolean or string, preventing potential bugs from unexpected data types in the ephemeralAgent state.	Low
General	Fail safe when probes inconclusive Relying on this.connectionState === 'connected' after a failed capability check can mask disconnections. Return false immediately when all probes fail to avoid stale connections being treated as healthy. packages/api/src/mcp/connection.ts [586-641] public async isConnected(): Promise<boolean> { - // First check if we're in a connected state if (this.connectionState !== 'connected') { return false; } try { - // Try ping first as it's the lightest check await this.client.ping(); - return this.connectionState === 'connected'; + return true; } catch (error) { - // Check if the error is because ping is not supported (method not found) const pingUnsupported = error instanceof Error && - ((error as Error)?.message.includes('-32601') || - (error as Error)?.message.includes('invalid method ping') || - (error as Error)?.message.includes('method not found')); + (error.message.includes('-32601') || + error.message.includes('invalid method ping') || + error.message.includes('method not found')); if (!pingUnsupported) { logger.error(`${this.getLogPrefix()} Ping failed:`, error); return false; } - // Ping is not supported by this server, try an alternative verification - logger.debug( - `${this.getLogPrefix()} Server does not support ping method, verifying connection with capabilities`, - ); + logger.debug(`${this.getLogPrefix()} Ping unsupported; verifying via capabilities`); try { - // Get server capabilities to verify connection is truly active const capabilities = this.client.getServerCapabilities(); - // If we have capabilities, try calling a supported method to verify connection if (capabilities?.tools) { await this.client.listTools(); - return this.connectionState === 'connected'; - } else if (capabilities?.resources) { + return true; + } + if (capabilities?.resources) { await this.client.listResources(); - return this.connectionState === 'connected'; - } else if (capabilities?.prompts) { + return true; + } + if (capabilities?.prompts) { await this.client.listPrompts(); - return this.connectionState === 'connected'; - } else { - // No capabilities to test, but we're in connected state and initialization succeeded - logger.debug( - `${this.getLogPrefix()} No capabilities to test, assuming connected based on state`, - ); - return this.connectionState === 'connected'; + return true; } + // No verifiable capability; treat as not connected to be safe + logger.warn(`${this.getLogPrefix()} No verifiable capability; treating as disconnected`); + return false; } catch (capabilityError) { - // If capability check fails, the connection is likely broken logger.error(`${this.getLogPrefix()} Connection verification failed:`, capabilityError); return false; } } } [To ensure code accuracy, apply this suggestion manually] Suggestion importance[1-10]: 6 __ Why: The suggestion correctly points out that assuming a connection is active when no verification method (ping or capability checks) is available can lead to false positives. Returning false in this ambiguous case is a safer, more robust approach to prevent using a potentially stale connection.	Low
More