feat(mcp): oauth#208
Merged
Merged
Conversation
…enhance logging - Removed redundant initialization call for MCP manager in the `register` function. - Added a new `waitInitialized` method to ensure MCP clients manager is initialized before operations. - Enhanced logging with a unique identifier for better traceability. - Introduced `MemoryMCPConfigStorage` for in-memory configuration management, including methods for saving, loading, and deleting configurations. - Added tests for `MemoryMCPConfigStorage` to ensure functionality and reliability.
…ance clarity - Revised the agent generation prompt to improve clarity and effectiveness. - Streamlined the process into clear steps, focusing on core intent extraction, persona design, and instruction architecture. - Added performance optimization strategies and output generation specifications. - Ensured all output content aligns with user language requests for better usability.
- Updated environment variables in `.env.example` and `README.md` to reflect the new Exa AI API. - Refactored web search components and tools to utilize Exa AI, including response handling and schema definitions. - Enhanced workflow examples to align with the new Exa API structure. - Adjusted node executor to support Exa AI tools for improved search capabilities.
…ization tools - Added a new `createTable` tool for generating interactive tables with sorting, filtering, and search capabilities. - Updated existing tools and documentation to reflect the new table functionality. - Enhanced the `RandomDataGeneratorExample` to include table creation in its workflow. - Modified various components to support the new table feature, including dynamic imports and UI adjustments. Co-authored-by: choi sung keun <86150470+cgoinglove@users.noreply.github.com>
… functionality - Updated the layout of the Interactive Table component to include additional padding and a more structured design. - Improved the search and export functionality by ensuring the dropdown menus for column visibility and export options are clearly defined. - Maintained existing features while enhancing the overall user experience and visual consistency.
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
Contributor
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Summary of Changes
Hello @cgoinglove, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request primarily focuses on enhancing the documentation related to the project's built-in data visualization tools. It provides a more comprehensive and detailed explanation of the interactive table features and chart generation capabilities. Note: The pull request title 'feat(mpc): oauth' appears to be a mismatch with the actual changes, which are purely documentation updates.
Highlights
- Documentation Enhancement: Updated
CLAUDE.mdandREADME.mdto provide a more detailed and structured description of the built-in data visualization tools. - Interactive Tables Features: Added comprehensive details about the advanced functionalities of interactive tables, including sorting, filtering, global search with highlighting, CSV/Excel export, column management, pagination, and data type support.
- Chart Generation Clarification: Clarified the capabilities of chart generation tools, specifically mentioning bar, line, and pie charts.
- Lazy-loaded Excel Export: Highlighted the implementation of lazy-loaded Excel export using a CDN-based
xlsxlibrary for optimal bundle size.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments or fill out our survey to provide feedback.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request introduces a new interactive table component and updates the documentation to reflect this and other data visualization capabilities. The pull request title should be updated to accurately reflect the changes, such as feat(ui): add interactive data table.
Comment on lines
+207
to
+212
| - **Interactive Tables**: Feature-rich tables with sorting, filtering, search, highlighting, pagination, and CSV/Excel export | ||
| - **Charts**: Generation of bar, line, and pie charts | ||
| - Tools generate parameters that render React components | ||
| - Components located in `@src/components/tool-invocation/` | ||
| - Dynamic imports used in `@src/components/message-parts.tsx` for rendering | ||
| - Lazy-loaded Excel export using CDN-based xlsx library for optimal bundle size |
There was a problem hiding this comment.
The detail about lazy-loading the Excel library is specific to the Interactive Tables feature. For better clarity and structure, I suggest nesting this information under the 'Interactive Tables' item. This makes it clear which feature benefits from this optimization, as it doesn't apply to the other visualization tools.
Suggested change
| - **Interactive Tables**: Feature-rich tables with sorting, filtering, search, highlighting, pagination, and CSV/Excel export | |
| - **Charts**: Generation of bar, line, and pie charts | |
| - Tools generate parameters that render React components | |
| - Components located in `@src/components/tool-invocation/` | |
| - Dynamic imports used in `@src/components/message-parts.tsx` for rendering | |
| - Lazy-loaded Excel export using CDN-based xlsx library for optimal bundle size | |
| - **Interactive Tables**: Feature-rich tables with sorting, filtering, search, highlighting, pagination, and CSV/Excel export | |
| - Lazy-loaded Excel export using CDN-based xlsx library for optimal bundle size |
| #### 📊 Data Visualization Tools | ||
|
|
||
| **Interactive Tables**: Create feature-rich data tables with advanced functionality: | ||
| - **Sorting & Filtering**: Sort by any column, filter data in real-time |
There was a problem hiding this comment.
The current implementation in interactive-table.tsx provides a global search that filters data, but not per-column filtering. To set accurate user expectations, it would be clearer to specify that filtering is done via a global search.
Suggested change
| - **Sorting & Filtering**: Sort by any column, filter data in real-time | |
| - **Sorting & Filtering**: Sort by any column, filter data in real-time using global search |
Contributor
|
Progress? @cgoinglove |
Collaborator
Author
|
@brrock 70% |
Contributor
|
Cool |
Collaborator
Author
|
The draft has been completed, but there are still parts that need to be tested and modified. The linear oauth at Since this is not a completed PR, there may be changes to the DB, so when testing this PR, please use a development database such as pnpm docker:pg. |
Collaborator
Author
|
/gemini summary |
Summary of ChangesI've added comprehensive OAuth 2.0 support for Model Context Protocol (MCP) servers, which allows for more secure and streamlined integration with external services. This feature includes new API routes to manage the OAuth flow and significant updates to the user interface to guide users through the authorization process. Additionally, I've refined the documentation to better describe the enhanced data visualization tools available. Highlights
Changelog
Activity
|
…provements - Updated the MCP page to sort the MCP list by status, prioritizing 'authorizing' status. - Reduced the refresh interval for the MCP list from 10 seconds to 5 seconds. - Improved the layout of the MCP card component for better content visibility and overflow handling. - Added a name property to the PgOAuthClientProvider for better logging context.
- Added logging for the count of MCP tools alongside MCP clients in the chat route. - Updated the state management in the useMcpList hook to filter allowed MCP servers based on the fetched data, improving data consistency.
- Moved the unlocking of the locker and tool info update to the finally block for better error management. - Added logging for tool info updates to enhance visibility into the MCPClient's operations.
brrock
approved these changes
Aug 4, 2025
…components - Enhanced logging in the chat route to handle undefined values for allowed tools, ensuring accurate logging output. - Updated the initial state in the store to include predefined allowed app default toolkits, improving state management. - Adjusted the store's partialization logic to filter allowed app default toolkits based on defined values, enhancing data integrity.
- Eliminated debug and info logging statements in the PgOAuthClientProvider class to streamline the code and reduce verbosity. - Maintained functionality while improving code clarity and maintainability.
commit 410ad4b Author: choi sung keun <86150470+cgoinglove@users.noreply.github.com> Date: Mon Aug 4 23:31:40 2025 +0900 refactor: remove unnecessary logging in PgOAuthClientProvider - Eliminated debug and info logging statements in the PgOAuthClientProvider class to streamline the code and reduce verbosity. - Maintained functionality while improving code clarity and maintainability. commit 3d003e0 Author: choi sung keun <86150470+cgoinglove@users.noreply.github.com> Date: Mon Aug 4 23:29:57 2025 +0900 refactor: improve logging and state initialization in chat and store components - Enhanced logging in the chat route to handle undefined values for allowed tools, ensuring accurate logging output. - Updated the initial state in the store to include predefined allowed app default toolkits, improving state management. - Adjusted the store's partialization logic to filter allowed app default toolkits based on defined values, enhancing data integrity. commit a656275 Author: choi sung keun <86150470+cgoinglove@users.noreply.github.com> Date: Mon Aug 4 22:56:54 2025 +0900 refactor: improve error handling and logging in MCPClient - Moved the unlocking of the locker and tool info update to the finally block for better error management. - Added logging for tool info updates to enhance visibility into the MCPClient's operations. commit d4c8ce1 Author: choi sung keun <86150470+cgoinglove@users.noreply.github.com> Date: Mon Aug 4 22:48:06 2025 +0900 refactor: enhance logging and state management in MCP components - Added logging for the count of MCP tools alongside MCP clients in the chat route. - Updated the state management in the useMcpList hook to filter allowed MCP servers based on the fetched data, improving data consistency. commit 487137e Merge: 7e31bf8 4dea9a7 Author: choi sung keun <86150470+cgoinglove@users.noreply.github.com> Date: Mon Aug 4 22:18:45 2025 +0900 Merge branch 'feat/mpc-oauth' into custom-main commit 7e31bf8 Author: cgoing <neo.cgoing@gmail.com> Date: Mon Aug 4 16:40:05 2025 +0900 refactor: update MCP client handling in OAuth callback - Changed the method to refresh the MCP client in the OAuth callback to ensure the latest client status is retrieved. - Updated the client addition logic in MCPClientsManager to await the completion of the addClient method, ensuring proper client management. commit 1b9eabf Author: cgoing <neo.cgoing@gmail.com> Date: Mon Aug 4 15:56:14 2025 +0900 fff commit 45e869e Author: cgoing <neo.cgoing@gmail.com> Date: Mon Aug 4 15:48:41 2025 +0900 Revert "asd" This reverts commit 2410dfd. commit 2410dfd Author: cgoing <neo.cgoing@gmail.com> Date: Mon Aug 4 15:41:15 2025 +0900 asd
- Changed the method for obtaining the MCP client in the OAuth callback from refreshClient to getClient to ensure accurate client status retrieval. - This adjustment enhances the reliability of client management within the OAuth flow.
- Replaced the refreshClient method with getClient in the OAuth callback to improve accuracy in client status retrieval. - This change enhances the reliability of client management within the OAuth flow.
- Introduced a new function to create OAuth response pages, improving code organization and readability. - Updated error and success response handling in the OAuth callback to utilize the new response creation function, ensuring consistent messaging and styling. - Improved logging for client refresh operations in the MCPClientsManager to provide clearer context during client updates.
- Updated the logging mechanism in the OAuth callback to provide clearer success and error messages. - Introduced a new logger configuration for consistent message formatting. - Enhanced the visibility of OAuth callback events for better debugging and monitoring.
…olSelectDropdown - Updated MCPCard and ToolSelectDropdown components to use ShieldAlertIcon instead of KeyRoundIcon for improved visual consistency. - Added authorization handling in ToolSelectDropdown, displaying an alert when authorization is required. - Adjusted connection timeout in create-mcp-client.ts for better performance in Vercel environment.
- Eliminated the logging of the authorization URL in the MCPClient class to reduce verbosity and improve code clarity. - Maintained the functionality of OAuth authorization handling while streamlining the logging process.
- Replaced the direct call to getOAuthSessionByState with a new method getAuthData to enhance code clarity and maintainability. - Ensured that the functionality of retrieving the OAuth code verifier remains intact while simplifying the implementation.
…-client.ts - Updated the connection timeout from 12000ms to 30000ms for improved performance in Vercel deployments. - This change aims to enhance the reliability of client connections under varying network conditions.
- Changed from Promise.allSettled to Promise.all for loading client configurations, ensuring that all client addition operations are awaited before proceeding. - This adjustment improves error handling and ensures that all clients are processed concurrently.
Collaborator
Author
|
/gemini review |
There was a problem hiding this comment.
Code Review
This pull request introduces OAuth support for MCP servers, which is a major and well-implemented feature. The changes span the entire stack, from new database tables for OAuth sessions to UI components that handle the authorization flow.
Overall, the implementation is robust, covering security considerations like using the state parameter to prevent CSRF and checking event.origin in postMessage handlers. The backend logic correctly handles the OAuth callback, token exchange, and updates the client status accordingly. The UI has also been thoughtfully updated to guide the user through the authorization process when required.
My main feedback focuses on improving maintainability by reducing code duplication and simplifying complex logic. I've identified a few areas where the code can be refactored for better clarity and long-term health.
Comment on lines
+85
to
+158
| const handleAuthorize = useCallback( | ||
| () => | ||
| pipeProcessing(() => | ||
| authorizeMcpClientAction(id).then((authUrl) => { | ||
| if (!authUrl) throw new Error("Not Authorizing"); | ||
| return new Promise((resolve, reject) => { | ||
| const authWindow = window.open( | ||
| authUrl, | ||
| "oauth", | ||
| "width=600,height=800,scrollbars=yes,resizable=yes", | ||
| ); | ||
| if (!authWindow) { | ||
| return reject( | ||
| new Error("Please allow popups for OAuth authentication"), | ||
| ); | ||
| } | ||
|
|
||
| let messageHandlerRegistered = false; | ||
| let intervalId: NodeJS.Timeout | null = null; | ||
|
|
||
| // Clean up function | ||
| const cleanup = () => { | ||
| if (messageHandlerRegistered) { | ||
| window.removeEventListener("message", messageHandler); | ||
| messageHandlerRegistered = false; | ||
| } | ||
| if (intervalId) { | ||
| clearInterval(intervalId); | ||
| intervalId = null; | ||
| } | ||
| }; | ||
|
|
||
| // Message handler for postMessage communication | ||
| const messageHandler = (event: MessageEvent) => { | ||
| // Security: only accept messages from same origin | ||
| if (event.origin !== window.location.origin) { | ||
| return; | ||
| } | ||
|
|
||
| if (event.data.type === "MCP_OAUTH_SUCCESS") { | ||
| cleanup(); | ||
| if (authWindow && !authWindow.closed) { | ||
| authWindow.close(); | ||
| } | ||
| resolve(true); | ||
| } else if (event.data.type === "MCP_OAUTH_ERROR") { | ||
| cleanup(); | ||
| if (authWindow && !authWindow.closed) { | ||
| authWindow.close(); | ||
| } | ||
| const errorMessage = | ||
| event.data.error_description || | ||
| event.data.error || | ||
| "Authentication failed"; | ||
| reject(new Error(errorMessage)); | ||
| } | ||
| }; | ||
|
|
||
| // Register message event listener | ||
| window.addEventListener("message", messageHandler); | ||
| messageHandlerRegistered = true; | ||
|
|
||
| // Backup: Poll for manual window close (in case postMessage fails) | ||
| intervalId = setInterval(() => { | ||
| if (authWindow.closed) { | ||
| cleanup(); | ||
| resolve(true); | ||
| } | ||
| }, 1000); | ||
| }); | ||
| }), | ||
| ), | ||
| [id], | ||
| ); |
There was a problem hiding this comment.
This handleAuthorize function contains complex logic for handling the OAuth popup window, including postMessage communication and a polling fallback. I've noticed this exact logic is duplicated in src/components/mcp-editor.tsx and src/components/tool-select-dropdown.tsx.
To improve maintainability and reduce code duplication, I strongly recommend extracting this logic into a reusable custom hook (e.g., useOAuthPopup) or a utility function. This would centralize the logic, make it easier to test, and ensure any future changes only need to be made in one place.
For example, a hook could look like this:
const { openAuthPopup, isAuthorizing } = useOAuthPopup();
const handleAuthorize = useCallback(() => {
pipeProcessing(() => openAuthPopup(id));
}, [id, openAuthPopup, pipeProcessing]);
Comment on lines
+50
to
+52
| ${Object.entries(postMessageData) | ||
| .map(([key, value]) => `${key}: '${value}'`) | ||
| .join(", ")} |
There was a problem hiding this comment.
The current method of constructing the postMessage data by mapping object entries to a string is a bit fragile. It assumes all values in postMessageData are strings and wraps them in single quotes. If a boolean or number is passed, it will be converted to a string (e.g., true becomes 'true').
While this works for the current implementation, a more robust approach would be to use JSON.stringify on the data object. This ensures that data types are preserved correctly and makes the code more resilient to future changes in postMessageData.
.map(([key, value]) => `${key}: ${JSON.stringify(value)})`)
.join(", ")}| await this.storage.init(this); | ||
| const configs = await this.storage.loadAll(); | ||
| await Promise.allSettled( | ||
| await Promise.all( |
There was a problem hiding this comment.
Using Promise.all here means that if any single client fails to initialize (e.g., due to a misconfiguration), the entire init process for the manager will fail.
Given that MCP servers are independent, it would be more resilient to use Promise.allSettled. This would allow the manager to initialize successfully even if some clients fail, logging the errors for later inspection without bringing down the whole system. The previous implementation used allSettled, and it seems more appropriate for this use case.
Suggested change
| await Promise.all( | |
| await Promise.allSettled( |
cgoinglove
pushed a commit
that referenced
this pull request
Aug 4, 2025
🤖 I have created a release *beep* *boop* --- ## [1.20.0](v1.19.1...v1.20.0) (2025-08-04) ### Features * add qwen3 coder to models file for openrouter ([#206](#206)) ([3731d00](3731d00)) * improve authentication configuration and social login handling ([#211](#211)) ([cd25937](cd25937)) * introduce interactive table creation and enhance visualization tools ([#205](#205)) ([623a736](623a736)) * **mcp:** oauth ([#208](#208)) ([136aded](136aded)) * **web-search:** replace Tavily API with Exa AI integration ([#204](#204)) ([7140487](7140487)) ### Bug Fixes * implement responsive horizontal layout for chat mention input with improved UX And generate Agent Prompt ([43ec980](43ec980)) * **mcp:** Safe MCP manager init logic for the Vercel environment ([#202](#202)) ([708fdfc](708fdfc)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
rockstarhub28
added a commit
to rockstarhub28/better-chatbot
that referenced
this pull request
Dec 24, 2025
🤖 I have created a release *beep* *boop* --- ## [1.20.0](keinsaasforever/better-chatbot@v1.19.1...v1.20.0) (2025-08-04) ### Features * add qwen3 coder to models file for openrouter ([#206](keinsaasforever/better-chatbot#206)) ([3731d00](keinsaasforever/better-chatbot@3731d00)) * improve authentication configuration and social login handling ([#211](keinsaasforever/better-chatbot#211)) ([cd25937](keinsaasforever/better-chatbot@cd25937)) * introduce interactive table creation and enhance visualization tools ([#205](keinsaasforever/better-chatbot#205)) ([623a736](keinsaasforever/better-chatbot@623a736)) * **mcp:** oauth ([#208](keinsaasforever/better-chatbot#208)) ([136aded](keinsaasforever/better-chatbot@136aded)) * **web-search:** replace Tavily API with Exa AI integration ([#204](keinsaasforever/better-chatbot#204)) ([7140487](keinsaasforever/better-chatbot@7140487)) ### Bug Fixes * implement responsive horizontal layout for chat mention input with improved UX And generate Agent Prompt ([43ec980](keinsaasforever/better-chatbot@43ec980)) * **mcp:** Safe MCP manager init logic for the Vercel environment ([#202](keinsaasforever/better-chatbot#202)) ([708fdfc](keinsaasforever/better-chatbot@708fdfc)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
CryptoWizardTeam
added a commit
to CryptoWizardTeam/ai-chatbot-2
that referenced
this pull request
Mar 14, 2026
🤖 I have created a release *beep* *boop* --- ## [1.20.0](keinsaasforever/better-chatbot@v1.19.1...v1.20.0) (2025-08-04) ### Features * add qwen3 coder to models file for openrouter ([#206](keinsaasforever/better-chatbot#206)) ([3731d00](keinsaasforever/better-chatbot@3731d00)) * improve authentication configuration and social login handling ([#211](keinsaasforever/better-chatbot#211)) ([cd25937](keinsaasforever/better-chatbot@cd25937)) * introduce interactive table creation and enhance visualization tools ([#205](keinsaasforever/better-chatbot#205)) ([623a736](keinsaasforever/better-chatbot@623a736)) * **mcp:** oauth ([#208](keinsaasforever/better-chatbot#208)) ([136aded](keinsaasforever/better-chatbot@136aded)) * **web-search:** replace Tavily API with Exa AI integration ([#204](keinsaasforever/better-chatbot#204)) ([7140487](keinsaasforever/better-chatbot@7140487)) ### Bug Fixes * implement responsive horizontal layout for chat mention input with improved UX And generate Agent Prompt ([43ec980](keinsaasforever/better-chatbot@43ec980)) * **mcp:** Safe MCP manager init logic for the Vercel environment ([#202](keinsaasforever/better-chatbot#202)) ([708fdfc](keinsaasforever/better-chatbot@708fdfc)) --- This PR was generated with [Release Please](https://github.com/googleapis/release-please). See [documentation](https://github.com/googleapis/release-please#release-please). Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.