fix(kb): fix mistral parse and kb uploads, include userId in internal auth

[waleedlatif1]

Summary

• fix mistral parse and kb uploads by including userId in knowledge-base context upload to workspace_files
• include userId in internal auth for enhanced security posture
• increment updated_at when inserting doc via knowledge block

Type of Change

• Bug fix

Testing

Tested extensively manually

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Preview	Comments	Updated (UTC)
docs	Skipped			Oct 30, 2025 6:16am

[greptile-apps]

Greptile Overview

Greptile Summary

This PR enhances the internal authentication system and fixes knowledge base document uploads by propagating user context through the entire processing pipeline.

Key Changes

• Internal Auth Enhancement: Modified generateInternalToken() and verifyInternalToken() to support optional userId in JWT payload, changing verification from returning a boolean to returning an object with {valid, userId?}
• KB Upload Fix: Added userId and workspaceId parameters throughout the document processing pipeline (processDocument, parseDocument, handleFileForOCR, etc.) to ensure proper metadata tracking when uploading files to storage
• Mistral Parser Fix: Fixed URL validation bug by converting relative paths to absolute URLs before validation, and included userId when generating internal tokens for API calls
• Updated Timestamp: Knowledge base updatedAt field now increments when documents are added via the knowledge block

Impact

All workflow routes now properly handle the new verification response structure, and knowledge base uploads will correctly track user and workspace ownership for files uploaded to cloud storage.

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The changes are well-structured and consistent throughout the codebase. The auth signature change is backwards-compatible (boolean to object with .valid property), all call sites have been updated, and the parameter threading through document processing maintains proper typing
• No files require special attention

Important Files Changed

File Analysis

Filename	Score	Overview
apps/sim/lib/auth/internal.ts	5/5	Enhanced JWT token generation and verification to include optional userId in payload for improved auth context
apps/sim/lib/auth/hybrid.ts	5/5	Updated hybrid auth to extract and use userId from internal JWT token payload, with fallback to query params and request body
apps/sim/lib/knowledge/documents/document-processor.ts	5/5	Added userId and workspaceId parameters throughout document processing pipeline to enable proper file metadata tracking
apps/sim/lib/knowledge/documents/service.ts	5/5	Updated knowledge base operations to increment updatedAt timestamp when documents are added, and pass user context to document processor
apps/sim/tools/mistral/parser.ts	5/5	Fixed URL validation by converting relative paths to absolute URLs before validation

Sequence Diagram

sequenceDiagram
    participant Client
    participant API as API Route
    participant Auth as Internal Auth
    participant DocProcessor as Document Processor
    participant Mistral as Mistral Parser
    participant Storage as Storage Service
    
    Note over Client,Storage: Enhanced Internal Auth Flow
    
    Client->>API: Request with Bearer token
    API->>Auth: Verify internal token
    Auth-->>API: Return validation result
    Note over API: Extract user from token payload
    
    Note over Client,Storage: Knowledge Base Upload Flow
    
    Client->>API: Upload document to KB
    API->>DocProcessor: Process with context parameters
    DocProcessor->>Storage: Upload file with metadata
    Note over Storage: Store metadata with context
    Storage-->>DocProcessor: Upload successful
    
    DocProcessor->>Mistral: Parse PDF if needed
    Note over Mistral: Detect internal route
    Mistral->>Auth: Request new token
    Auth-->>Mistral: Provide token
    Mistral->>Mistral: Call internal API
    Note over Mistral: Context preserved in chain
    Mistral-->>DocProcessor: Return parsed content
    
    DocProcessor-->>API: Processing complete
    API-->>Client: Success

Loading

_{7 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}