refactor: major cleanup of legacy and temporary files
- Removed ~965MB of temporary directories (temp/, extracted-nodes/, etc) - Deleted outdated database files and backups (.bak files) - Removed legacy shell scripts (mcp-server-v20.sh, rebuild-v20.sh) - Cleaned up orphan test files and debugging scripts - Removed duplicate schema file (src/db/schema.sql) - Deleted old Dockerfile.old and empty database files - Updated documentation structure in README.md - Added n8n_diagnostic tool to documentation - Condensed version history in CLAUDE.md - Created release notes for v2.7.0 Total impact: Removed 34 files, saving ~965MB+ of disk space 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
czlonkowski
@@ -1,424 +0,0 @@
|
||||
# Changelog
|
||||
|
||||
All notable changes to this project will be documented in this file.
|
||||
|
||||
## [2.7.0] - 2025-06-28
|
||||
|
||||
### Added
|
||||
- **Diff-Based Workflow Editing**: Revolutionary token-efficient workflow updates
|
||||
- **NEW**: `n8n_update_partial_workflow` tool - Update workflows using diff operations for precise, incremental changes
|
||||
- **NEW**: WorkflowDiffEngine - Applies targeted edits without sending full workflow JSON
|
||||
- **NEW**: 13 diff operations - addNode, removeNode, updateNode, moveNode, enableNode, disableNode, addConnection, removeConnection, updateConnection, updateSettings, updateName, addTag, removeTag
|
||||
- **NEW**: Transaction safety - Validates all operations before applying any changes
|
||||
- **NEW**: Validation-only mode - Test your diff operations without applying them
|
||||
- **NEW**: Transactional Updates - Two-pass processing allows adding nodes and connections in any order
|
||||
- Smart node references - Use either node ID or name for operations
|
||||
- Order independence - Add connections before nodes, engine handles dependencies automatically
|
||||
- Operation limit - Maximum 5 operations per request ensures reliability
|
||||
- 80-90% token savings - Only send the changes, not the entire workflow
|
||||
- Comprehensive test coverage - All operations and edge cases tested
|
||||
- Example guide - See [workflow-diff-examples.md](./workflow-diff-examples.md) for usage patterns
|
||||
|
||||
### Changed
|
||||
- **RENAMED**: `n8n_update_workflow` → `n8n_update_full_workflow` - Clarifies that it replaces the entire workflow
|
||||
|
||||
### Fixed
|
||||
- **Docker Runtime Dependencies**: Added missing `uuid` package to runtime dependencies
|
||||
- Fixed MODULE_NOT_FOUND error in Docker containers
|
||||
- Updated both package.json and package.runtime.json
|
||||
- Aligned package versions between main and runtime configurations
|
||||
|
||||
## [2.6.2] - 2025-06-26
|
||||
|
||||
### Added
|
||||
- **Enhanced Workflow Creation Validation**: Major improvements to prevent broken workflows
|
||||
- **NEW**: Node type existence validation - Verifies node types actually exist in n8n
|
||||
- **NEW**: Smart suggestions for common mistakes - Detects `webhook` and suggests `n8n-nodes-base.webhook`
|
||||
- **NEW**: Minimum viable workflow validation - Prevents single-node workflows (except webhooks)
|
||||
- **NEW**: Empty connection detection - Catches multi-node workflows with no connections
|
||||
- **NEW**: Helper functions - `getWorkflowStructureExample()` and `getWorkflowFixSuggestions()`
|
||||
- Enhanced error messages with clear guidance and connection examples
|
||||
- Prevents broken workflows that show as question marks in n8n UI
|
||||
|
||||
### Fixed
|
||||
- **Critical**: `nodes-base.webhook` validation - Now caught BEFORE database lookup
|
||||
- Previously, `nodes-base.webhook` would pass validation because it existed in the normalized database
|
||||
- Now explicitly checks for and rejects `nodes-base.` prefix before any database operations
|
||||
- This fixes the exact issue causing workflows to appear broken in n8n UI
|
||||
|
||||
### Changed
|
||||
- Workflow validator now checks node types in order: invalid patterns → database lookup → suggestions
|
||||
- Connection validation messages now include proper format examples
|
||||
- Error messages are more actionable with specific fix instructions
|
||||
|
||||
## [2.6.1] - 2025-06-26
|
||||
|
||||
### Added
|
||||
- **Enhanced typeVersion Validation**: Comprehensive validation for versioned nodes
|
||||
- **NEW**: typeVersion validation enforced on all versioned nodes
|
||||
- Catches missing typeVersion and provides correct version to use
|
||||
- Warns on outdated versions when newer versions are available
|
||||
- Prevents invalid versions that exceed maximum supported
|
||||
- Helps AI agents avoid common workflow creation mistakes
|
||||
- Test coverage for all typeVersion scenarios
|
||||
|
||||
### Changed
|
||||
- WorkflowValidator now includes typeVersion in its validation pipeline
|
||||
- Enhanced error messages for typeVersion issues with actionable suggestions
|
||||
|
||||
## [2.6.0] - 2025-06-26
|
||||
|
||||
### Added
|
||||
- **n8n Management Tools Integration**: Complete workflow lifecycle management via API
|
||||
- **NEW**: 14 n8n management tools for creating, updating, and executing workflows
|
||||
- **NEW**: `n8n_create_workflow` - Create workflows programmatically with validation
|
||||
- **NEW**: `n8n_update_workflow` - Update existing workflows with safety checks
|
||||
- **NEW**: `n8n_trigger_webhook_workflow` - Execute workflows via webhooks
|
||||
- **NEW**: `n8n_list_executions` - Monitor workflow executions with filtering
|
||||
- **NEW**: `n8n_health_check` - Check n8n instance connectivity and features
|
||||
- **NEW**: Workflow management tools with smart error handling
|
||||
- **NEW**: Execution management tools for monitoring and debugging
|
||||
- Integrated n8n-manager-for-ai-agents functionality
|
||||
- Optional feature - only enabled when N8N_API_URL and N8N_API_KEY configured
|
||||
- Complete workflow lifecycle: discover → build → validate → deploy → execute
|
||||
- Smart error handling for API limitations (activation, direct execution)
|
||||
- Conditional tool registration based on configuration
|
||||
|
||||
### Changed
|
||||
- Updated `start_here_workflow_guide` to include n8n management tools documentation
|
||||
- Enhanced MCP server to conditionally register management tools
|
||||
- Added comprehensive integration tests for all management features
|
||||
|
||||
## [2.5.1] - 2025-01-25
|
||||
|
||||
### Added
|
||||
- **AI Tool Support Enhancement**: Major improvements to AI tool integration
|
||||
- **NEW**: `get_node_as_tool_info` tool - Get specific information about using ANY node as an AI tool
|
||||
- **ENHANCED**: `get_node_info` now includes `aiToolCapabilities` section for all nodes
|
||||
- **ENHANCED**: `list_ai_tools` - Added usage guidance explaining ANY node can be used as a tool
|
||||
- **ENHANCED**: `WorkflowValidator` - Now validates `ai_tool` connections in workflows
|
||||
- AI workflow pattern detection - Warns when AI Agents have no tools connected
|
||||
- Community node detection - Reminds about N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE environment variable
|
||||
- **NEW**: AI Tool TaskTemplates - Added use_google_sheets_as_tool, use_slack_as_tool, multi_tool_ai_agent
|
||||
- Comprehensive examples showing how to connect regular nodes as AI tools
|
||||
- Tool usage documentation with $fromAI() expression examples
|
||||
|
||||
### Changed
|
||||
- Clarified that ANY node can be used as an AI tool, not just nodes with usableAsTool property
|
||||
- Enhanced workflow validation to understand and validate ai_tool connections
|
||||
- Improved expression validation to support $fromAI() dynamic AI parameters
|
||||
|
||||
## [2.5.0] - 2025-01-20
|
||||
|
||||
### Added
|
||||
- **Complete Workflow Validation**: Professional-grade workflow validation system
|
||||
- **NEW**: `validate_workflow` tool - Validate entire workflows before deployment
|
||||
- **NEW**: `validate_workflow_connections` tool - Check workflow structure and connections
|
||||
- **NEW**: `validate_workflow_expressions` tool - Validate all n8n expressions in a workflow
|
||||
- **NEW**: `ExpressionValidator` - Comprehensive n8n expression syntax validation
|
||||
- **NEW**: `WorkflowValidator` - Complete workflow structure and logic validation
|
||||
- Detects cycles (infinite loops) in workflows
|
||||
- Validates node references in expressions ($node["Node Name"])
|
||||
- Checks for orphaned nodes and missing connections
|
||||
- Expression syntax validation with common mistake detection
|
||||
- Workflow best practices analysis with suggestions
|
||||
- Supports partial validation (nodes only, connections only, expressions only)
|
||||
- Test coverage for all validation scenarios
|
||||
|
||||
## [2.4.2] - 2025-01-15
|
||||
|
||||
### Added
|
||||
- **Enhanced Node Configuration Validation**: Operation-aware validation with dramatic accuracy improvements
|
||||
- **NEW**: `validate_node_operation` tool - Operation-aware validation with 80%+ fewer false positives
|
||||
- **NEW**: `validate_node_minimal` tool - Lightning-fast validation for just required fields
|
||||
- **NEW**: Validation profiles - Choose between minimal, runtime, ai-friendly, or strict validation
|
||||
- **NEW**: `EnhancedConfigValidator` - Smart validation that only checks relevant properties
|
||||
- **NEW**: Node-specific validators - Custom logic for Slack, Google Sheets, OpenAI, MongoDB, Webhook, Postgres, MySQL
|
||||
- **NEW**: SQL safety features - Detects SQL injection risks, unsafe DELETE/UPDATE queries
|
||||
- Added operation context filtering (only validates properties for selected operation)
|
||||
- Integrated working examples in validation responses when errors found
|
||||
- Added actionable next steps and auto-fix suggestions
|
||||
- Basic code syntax validation for JavaScript/Python in Code node
|
||||
- Dramatic improvement for complex multi-operation nodes
|
||||
- Test results: Slack validation reduced from 45 errors to 1 error!
|
||||
|
||||
### Removed
|
||||
- Deprecated `validate_node_config` tool in favor of new operation-aware validators
|
||||
|
||||
## [2.4.1] - 2025-01-10
|
||||
|
||||
### Added
|
||||
- **n8n Workflow Templates**: Integration with n8n.io template library
|
||||
- **NEW**: `list_node_templates` tool - Find workflow templates using specific nodes
|
||||
- **NEW**: `get_template` tool - Get complete workflow JSON for import
|
||||
- **NEW**: `search_templates` tool - Search templates by keywords
|
||||
- **NEW**: `get_templates_for_task` tool - Get curated templates for common tasks
|
||||
- Added Templates system with n8n.io API integration
|
||||
- Templates filtered to last 6 months only (freshness guarantee)
|
||||
- Manual fetch system - not part of regular rebuild
|
||||
- Full workflow JSON available for immediate use
|
||||
- 10 task categories: AI automation, data sync, webhooks, etc.
|
||||
|
||||
## [2.4.0] - 2025-01-05
|
||||
|
||||
### Added
|
||||
- **AI-Optimized MCP Tools**: Dramatically improved AI agent experience
|
||||
- **NEW**: `get_node_essentials` tool - Returns only 10-20 essential properties (95% size reduction)
|
||||
- **NEW**: `search_node_properties` tool - Search for specific properties within nodes
|
||||
- **NEW**: `get_node_for_task` tool - Pre-configured settings for 14 common tasks
|
||||
- **NEW**: `list_tasks` tool - Discover available task templates
|
||||
- **NEW**: `validate_node_config` tool - Validate configurations before use
|
||||
- **NEW**: `get_property_dependencies` tool - Analyze property visibility dependencies
|
||||
- Added PropertyFilter service with curated essential properties for 20+ nodes
|
||||
- Added ExampleGenerator with working examples for common use cases
|
||||
- Added TaskTemplates service with 14 pre-configured tasks
|
||||
- Added ConfigValidator service for comprehensive validation
|
||||
- Added PropertyDependencies service for dependency analysis
|
||||
- Enhanced all property descriptions - 100% coverage
|
||||
- Added version information to essentials response
|
||||
- Response sizes reduced from 100KB+ to <5KB for common nodes
|
||||
|
||||
### Changed
|
||||
- **License Change**: Changed from Apache 2.0 to MIT License for wider adoption
|
||||
- Fixed missing AI and LangChain node documentation
|
||||
- Improved documentation mapping for better coverage
|
||||
|
||||
## [2.3.3] - 2025-06-16
|
||||
|
||||
### Added
|
||||
- **Automated Dependency Update System**: Comprehensive solution for keeping n8n packages in sync
|
||||
- Custom update script (`scripts/update-n8n-deps.js`) that respects n8n's interdependencies
|
||||
- GitHub Actions workflow for weekly automated updates
|
||||
- Renovate configuration as an alternative solution
|
||||
- Dependency update documentation guide
|
||||
- Support for automatic n8n package version synchronization
|
||||
- Documentation updates reflecting current metrics
|
||||
|
||||
### Fixed
|
||||
- **Validation Script Node Type References**: Fixed node type format issues
|
||||
- Changed from short names (e.g., 'httpRequest') to full names (e.g., 'nodes-base.httpRequest')
|
||||
- Removed versioned check for Code node as it's not consistently detected
|
||||
- All validation tests now pass after dependency updates
|
||||
|
||||
### Changed
|
||||
- Updated n8n dependencies to latest versions:
|
||||
- n8n: 1.14.1 → 1.97.1
|
||||
- n8n-core: 1.14.1 → 1.96.0
|
||||
- n8n-workflow: 1.82.0 → 1.94.0
|
||||
- @n8n/n8n-nodes-langchain: 1.97.0 → 1.96.1
|
||||
- Significant increase in detected nodes and capabilities:
|
||||
- Total nodes: 458 → 525
|
||||
- AI-capable tools: 35 → 263 (major increase due to updated detection)
|
||||
- Nodes with properties: 98.7% → 99%
|
||||
- Nodes with operations: 57.9% → 63.6%
|
||||
|
||||
### Technical Details
|
||||
- Dependency update script now checks n8n's required dependency versions
|
||||
- Validation script uses correct database column names
|
||||
- All critical nodes (httpRequest, code, slack, agent) validate successfully
|
||||
|
||||
## [2.3.2] - 2025-06-14
|
||||
|
||||
### Fixed
|
||||
- **HTTP Server Stream Error**: Complete fix for "stream is not readable" error
|
||||
- Removed Express body parsing middleware that was consuming request streams
|
||||
- Fixed "Server not initialized" error with direct JSON-RPC implementation
|
||||
- Added `USE_FIXED_HTTP=true` environment variable for stable HTTP mode
|
||||
- Bypassed problematic StreamableHTTPServerTransport implementation
|
||||
- HTTP server now works reliably with average response time of ~12ms
|
||||
- Updated all HTTP server implementations to preserve raw streams
|
||||
|
||||
### Added
|
||||
- `http-server-fixed.ts` - Direct JSON-RPC implementation
|
||||
- `ConsoleManager` utility for stream isolation
|
||||
- `MCP Engine` interface for service integration
|
||||
- Comprehensive documentation for HTTP server fixes
|
||||
|
||||
### Changed
|
||||
- Default HTTP mode now uses fixed implementation when `USE_FIXED_HTTP=true`
|
||||
- Updated Docker configuration to use fixed implementation by default
|
||||
- Improved error handling and logging in HTTP mode
|
||||
|
||||
## [2.3.1] - 2025-06-14
|
||||
|
||||
### Added
|
||||
- **Single-Session Architecture**: Initial attempt to fix HTTP server issues
|
||||
- Implemented session reuse across requests
|
||||
- Added console output isolation
|
||||
- Created engine interface for service integration
|
||||
|
||||
### Fixed
|
||||
- Partial fix for "stream is not readable" error (completed in v2.3.2)
|
||||
|
||||
## [2.3.0] - 2024-12-06
|
||||
|
||||
### Added
|
||||
- **HTTP Remote Deployment**: Single-user HTTP server for remote access
|
||||
- Stateless architecture for simple deployments
|
||||
- Bearer token authentication
|
||||
- Compatible with mcp-remote adapter for Claude Desktop
|
||||
- New HTTP mode scripts and deployment helper
|
||||
- **Universal Node.js Compatibility**: Automatic database adapter fallback system
|
||||
- Primary adapter: `better-sqlite3` for optimal performance
|
||||
- Fallback adapter: `sql.js` (pure JavaScript) for version mismatches
|
||||
- Automatic detection and switching between adapters
|
||||
- No manual configuration required
|
||||
- Database adapter abstraction layer (`src/database/database-adapter.ts`)
|
||||
- Version detection and logging for troubleshooting
|
||||
- sql.js dependency for pure JavaScript SQLite implementation
|
||||
- HTTP server implementation (`src/http-server.ts`)
|
||||
- Deployment documentation and scripts
|
||||
|
||||
### Changed
|
||||
- Updated all database operations to use the adapter interface
|
||||
- Removed Node.js v20.17.0 requirement - now works with ANY version
|
||||
- Simplified Claude Desktop setup - no wrapper scripts needed
|
||||
- Enhanced error messages for database initialization
|
||||
- Made all MCP tool handlers async for proper initialization
|
||||
|
||||
### Fixed
|
||||
- NODE_MODULE_VERSION mismatch errors with Claude Desktop
|
||||
- Native module compilation issues in restricted environments
|
||||
- Compatibility issues when running with different Node.js versions
|
||||
- Database initialization race conditions in HTTP mode
|
||||
|
||||
### Technical Details
|
||||
- Better-sqlite3: ~10-50x faster (when compatible)
|
||||
- sql.js: ~2-5x slower but universally compatible
|
||||
- Both adapters maintain identical API and functionality
|
||||
- Automatic persistence for sql.js with 100ms debounced saves
|
||||
- HTTP server uses StreamableHTTPServerTransport for MCP compatibility
|
||||
|
||||
## [2.2.0] - 2024-12-06
|
||||
|
||||
### Added
|
||||
- PropertyExtractor class for dedicated property/operation extraction
|
||||
- NodeRepository for proper JSON serialization/deserialization
|
||||
- Support for @n8n/n8n-nodes-langchain package (59 AI nodes)
|
||||
- AI tool detection (35 tools with usableAsTool property)
|
||||
- Test suite for critical node validation
|
||||
- Comprehensive documentation (README, SETUP, CHANGELOG)
|
||||
- Example configuration files for Claude Desktop
|
||||
- Node.js v20.17.0 wrapper scripts for compatibility
|
||||
|
||||
### Fixed
|
||||
- Empty properties/operations arrays (now 98.7% nodes have properties)
|
||||
- Versioned node detection (HTTPRequest, Code properly identified)
|
||||
- Documentation mapping for nodes with directory-based docs
|
||||
- Critical node validation (httpRequest, slack, code all pass)
|
||||
|
||||
### Changed
|
||||
- Refactored parser to handle instance-level properties
|
||||
- Updated MCP server to use NodeRepository
|
||||
- Improved rebuild script with validation
|
||||
- Enhanced database schema with proper typing
|
||||
|
||||
### Metrics
|
||||
- 458 total nodes (100% success rate)
|
||||
- 452 nodes with properties (98.7%)
|
||||
- 265 nodes with operations (57.9%)
|
||||
- 406 nodes with documentation (88.6%)
|
||||
- 35 AI-capable tools detected
|
||||
- All critical nodes validated
|
||||
|
||||
## [2.1.0] - 2025-01-08
|
||||
|
||||
### Added
|
||||
- Remote deployment capabilities via HTTP/JSON-RPC transport
|
||||
- Domain configuration through environment variables (`MCP_DOMAIN`)
|
||||
- Bearer token authentication for remote access
|
||||
- Comprehensive remote deployment documentation
|
||||
- PM2 and Nginx configuration examples
|
||||
- HTTP server mode (`npm run start:http`)
|
||||
|
||||
### Enhanced
|
||||
- Support for both local (stdio) and remote (HTTP) deployment modes
|
||||
- Production deployment guide for VM/cloud environments
|
||||
- Claude Desktop configuration for remote servers
|
||||
|
||||
## [2.0.0] - 2025-01-08
|
||||
|
||||
### Major Refactoring
|
||||
- **BREAKING CHANGE**: Refocused project to serve only n8n node documentation
|
||||
- Removed all workflow execution and management features
|
||||
- Removed bidirectional n8n-MCP integration
|
||||
- Simplified to be a read-only documentation server
|
||||
|
||||
### Added
|
||||
- SQLite database with full-text search (FTS5) for node documentation
|
||||
- Integration with n8n-docs repository for official documentation
|
||||
- Automatic example workflow generation for each node type
|
||||
- Comprehensive node information including:
|
||||
- Source code
|
||||
- Official documentation
|
||||
- Usage examples
|
||||
- Properties schema
|
||||
- Credential definitions
|
||||
|
||||
### New MCP Tools
|
||||
- `list_nodes` - List available nodes with filtering
|
||||
- `get_node_info` - Get complete node information
|
||||
- `search_nodes` - Full-text search across nodes
|
||||
- `get_node_example` - Get example workflow for a node
|
||||
- `get_node_source_code` - Get only source code
|
||||
- `get_node_documentation` - Get only documentation
|
||||
- `rebuild_database` - Rebuild entire node database
|
||||
- `get_database_statistics` - Database statistics
|
||||
|
||||
### Infrastructure
|
||||
- New database schema optimized for documentation storage
|
||||
- `DocumentationFetcher` for n8n-docs repository integration
|
||||
- `ExampleGenerator` for creating node usage examples
|
||||
- `NodeDocumentationService` for database management
|
||||
|
||||
## [1.1.0] - 2024-01-07
|
||||
|
||||
### Added
|
||||
- Node source code extraction capability via `get_node_source_code` tool
|
||||
- List available nodes functionality with `list_available_nodes` tool
|
||||
- `NodeSourceExtractor` utility for file system access to n8n nodes
|
||||
- Resource endpoint `nodes://source/{nodeType}` for accessing node source code
|
||||
- Docker test environment with mounted n8n node_modules
|
||||
- Comprehensive test suite for AI Agent node extraction
|
||||
- Test documentation in `docs/AI_AGENT_EXTRACTION_TEST.md`
|
||||
|
||||
### Enhanced
|
||||
- MCP server can now access and extract n8n node implementations
|
||||
- Support for extracting credential definitions alongside node code
|
||||
- Package metadata included in extraction results
|
||||
|
||||
## [1.0.0] - 2024-01-07
|
||||
|
||||
### Initial Release
|
||||
- Complete n8n-MCP integration implementation
|
||||
- MCP server exposing n8n workflows as tools, resources, and prompts
|
||||
- Custom n8n node for connecting to MCP servers
|
||||
- Bidirectional data format conversion bridge
|
||||
- Token-based authentication system
|
||||
- Comprehensive error handling and logging
|
||||
- Full test coverage for core components
|
||||
- Docker support with production and development configurations
|
||||
- Installation scripts for n8n custom node deployment
|
||||
|
||||
### MCP Tools
|
||||
- `execute_workflow` - Execute n8n workflows
|
||||
- `list_workflows` - List available workflows
|
||||
- `get_workflow` - Get workflow details
|
||||
- `create_workflow` - Create new workflows
|
||||
- `update_workflow` - Update existing workflows
|
||||
- `delete_workflow` - Delete workflows
|
||||
- `get_executions` - Get execution history
|
||||
- `get_execution_data` - Get execution details
|
||||
|
||||
### MCP Resources
|
||||
- `workflow://active` - Active workflows
|
||||
- `workflow://all` - All workflows
|
||||
- `execution://recent` - Recent executions
|
||||
- `credentials://types` - Credential types
|
||||
- `nodes://available` - Available nodes
|
||||
|
||||
### MCP Prompts
|
||||
- `create_workflow_prompt` - Workflow creation
|
||||
- `debug_workflow_prompt` - Workflow debugging
|
||||
- `optimize_workflow_prompt` - Workflow optimization
|
||||
- `explain_workflow_prompt` - Workflow explanation
|
||||
@@ -1,163 +0,0 @@
|
||||
# Docker stdio Fix Implementation Plan for n8n-MCP
|
||||
|
||||
Based on community research and successful MCP Docker deployments, here's a streamlined fix for the initialization timeout issue.
|
||||
|
||||
## Root Cause
|
||||
|
||||
Docker treats container stdout as a pipe (not TTY), causing block buffering. The MCP server's JSON-RPC responses sit in the buffer instead of being immediately sent to Claude Desktop, causing a 60-second timeout.
|
||||
|
||||
## Implementation Steps
|
||||
|
||||
### Step 1: Test Simple Interactive Mode
|
||||
|
||||
First, verify if just using `-i` flag solves the issue:
|
||||
|
||||
**Update README.md:**
|
||||
```json
|
||||
{
|
||||
"mcpServers": {
|
||||
"n8n-mcp": {
|
||||
"command": "docker",
|
||||
"args": [
|
||||
"run",
|
||||
"-i", // Interactive mode - keeps stdin open
|
||||
"--rm",
|
||||
"-e", "MCP_MODE=stdio",
|
||||
"-e", "LOG_LEVEL=error",
|
||||
"-e", "DISABLE_CONSOLE_OUTPUT=true",
|
||||
"ghcr.io/czlonkowski/n8n-mcp:latest"
|
||||
]
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
**Test command:**
|
||||
```bash
|
||||
echo '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05"},"id":1}' | \
|
||||
docker run -i --rm \
|
||||
-e MCP_MODE=stdio \
|
||||
-e LOG_LEVEL=error \
|
||||
-e DISABLE_CONSOLE_OUTPUT=true \
|
||||
ghcr.io/czlonkowski/n8n-mcp:latest
|
||||
```
|
||||
|
||||
Expected: Should receive a JSON response immediately.
|
||||
|
||||
### Step 2: Add Explicit Stdout Flushing (If Needed)
|
||||
|
||||
If Step 1 doesn't work, add minimal flushing to the Node.js server:
|
||||
|
||||
**File: `src/mcp/server-update.ts`**
|
||||
|
||||
Update the `run()` method:
|
||||
```typescript
|
||||
async run(): Promise<void> {
|
||||
await this.ensureInitialized();
|
||||
|
||||
const transport = new StdioServerTransport();
|
||||
await this.server.connect(transport);
|
||||
|
||||
// Ensure stdout is not buffered in Docker
|
||||
if (!process.stdout.isTTY && process.env.IS_DOCKER) {
|
||||
// Force unbuffered stdout
|
||||
process.stdout.write('');
|
||||
}
|
||||
|
||||
logger.info('n8n Documentation MCP Server running on stdio transport');
|
||||
|
||||
// Keep process alive
|
||||
process.stdin.resume();
|
||||
}
|
||||
```
|
||||
|
||||
**File: `Dockerfile`**
|
||||
|
||||
Add environment variable:
|
||||
```dockerfile
|
||||
ENV IS_DOCKER=true
|
||||
```
|
||||
|
||||
### Step 3: System-Level Unbuffering (Last Resort)
|
||||
|
||||
Only if Steps 1-2 fail, implement stdbuf wrapper:
|
||||
|
||||
**File: `docker-entrypoint.sh`**
|
||||
```bash
|
||||
#!/bin/sh
|
||||
# Force line buffering for stdio communication
|
||||
exec stdbuf -oL -eL node /app/dist/mcp/index.js
|
||||
```
|
||||
|
||||
**File: `Dockerfile`**
|
||||
```dockerfile
|
||||
# Add stdbuf utility
|
||||
RUN apk add --no-cache coreutils
|
||||
|
||||
# Copy and setup entrypoint
|
||||
COPY docker-entrypoint.sh /docker-entrypoint.sh
|
||||
RUN chmod +x /docker-entrypoint.sh
|
||||
|
||||
ENTRYPOINT ["/docker-entrypoint.sh"]
|
||||
```
|
||||
|
||||
## Testing Protocol
|
||||
|
||||
### 1. Local Docker Test
|
||||
```bash
|
||||
# Build test image
|
||||
docker build -t n8n-mcp:test .
|
||||
|
||||
# Test with echo pipe
|
||||
echo '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05"},"id":1}' | \
|
||||
docker run -i --rm n8n-mcp:test | head -1
|
||||
|
||||
# Should see immediate JSON response
|
||||
```
|
||||
|
||||
### 2. Claude Desktop Test
|
||||
1. Update `claude_desktop_config.json` with new configuration
|
||||
2. Restart Claude Desktop
|
||||
3. Check Developer tab for "running" status
|
||||
4. Test a simple MCP command
|
||||
|
||||
### 3. Debug if Needed
|
||||
```bash
|
||||
# Run with stderr output for debugging
|
||||
docker run -i --rm \
|
||||
-e MCP_MODE=stdio \
|
||||
-e LOG_LEVEL=debug \
|
||||
ghcr.io/czlonkowski/n8n-mcp:latest 2>debug.log
|
||||
```
|
||||
|
||||
## Success Criteria
|
||||
|
||||
- [ ] No timeout errors in Claude Desktop logs
|
||||
- [ ] MCP tools are accessible immediately
|
||||
- [ ] No "Shutting down..." messages in stdout
|
||||
- [ ] Simple `-i` flag configuration works
|
||||
|
||||
## Rollout Plan
|
||||
|
||||
1. **Test locally** with simple `-i` flag first
|
||||
2. **Update Docker image** only if code changes needed
|
||||
3. **Update README** with working configuration
|
||||
4. **Community announcement** with simple Docker instructions
|
||||
|
||||
## Key Insights from Research
|
||||
|
||||
- Most MCP Docker deployments work with just `-i` flag
|
||||
- Complex solutions often unnecessary
|
||||
- Node.js typically doesn't need explicit unbuffering (unlike Python with `-u`)
|
||||
- Claude Desktop only supports stdio for local servers (not HTTP)
|
||||
- Proper testing can quickly identify if buffering is the actual issue
|
||||
|
||||
## What NOT to Do
|
||||
|
||||
- Don't add TTY flag (`-t`) - it's for terminal UI, not needed for MCP
|
||||
- Don't implement complex multi-phase solutions
|
||||
- Don't switch to HTTP transport (Claude Desktop doesn't support it locally)
|
||||
- Don't modify MCP protocol handling
|
||||
- Don't add unnecessary wrapper scripts unless proven necessary
|
||||
|
||||
The solution should be as simple as possible - likely just the `-i` flag in the Docker command.
|
||||
@@ -1,252 +0,0 @@
|
||||
# Docker MCP Initialization Timeout Fix Plan
|
||||
|
||||
## Problem Summary
|
||||
|
||||
The n8n-MCP Docker container fails to work with Claude Desktop due to MCP initialization timeout:
|
||||
|
||||
1. Claude sends `initialize` request
|
||||
2. Server receives it (logs show "Message from client: {"method":"initialize"...}")
|
||||
3. Server appears to connect successfully
|
||||
4. **No response is sent back to Claude**
|
||||
5. Claude times out after 60 seconds
|
||||
6. Container outputs "Shutting down..." which breaks JSON-RPC protocol
|
||||
|
||||
## Root Cause Analysis
|
||||
|
||||
### 1. **Stdout Buffering in Docker**
|
||||
|
||||
Docker containers often buffer stdout, especially when not running with TTY (`-t` flag). This is the most likely culprit:
|
||||
|
||||
- Node.js/JavaScript may buffer stdout when not connected to a TTY
|
||||
- Docker's stdout handling differs from direct execution
|
||||
- The MCP response might be stuck in the buffer
|
||||
|
||||
**Evidence:**
|
||||
- Common Docker issue (moby/moby#1385, docker/compose#1549)
|
||||
- Python requires `-u` flag for unbuffered output in Docker
|
||||
- Different base images have different buffering behavior
|
||||
|
||||
### 2. **MCP SDK Server Connection Issue**
|
||||
|
||||
The server might not be properly completing the connection handshake:
|
||||
|
||||
```typescript
|
||||
const transport = new StdioServerTransport();
|
||||
await server.connect(transport); // This might not complete properly
|
||||
```
|
||||
|
||||
### 3. **Missing Initialize Handler**
|
||||
|
||||
While the MCP SDK should handle `initialize` automatically, there might be an issue with:
|
||||
- Handler registration order
|
||||
- Server capabilities configuration
|
||||
- Transport initialization timing
|
||||
|
||||
### 4. **Process Lifecycle Management**
|
||||
|
||||
The container might be:
|
||||
- Exiting too early
|
||||
- Not keeping the event loop alive
|
||||
- Missing proper signal handling
|
||||
|
||||
## Fixing Plan
|
||||
|
||||
### Phase 1: Immediate Fixes (High Priority)
|
||||
|
||||
#### 1.1 Force Stdout Flushing
|
||||
|
||||
**File:** `src/mcp/server-update.ts`
|
||||
|
||||
Add explicit stdout flushing after server connection:
|
||||
|
||||
```typescript
|
||||
async run(): Promise<void> {
|
||||
await this.ensureInitialized();
|
||||
|
||||
const transport = new StdioServerTransport();
|
||||
await this.server.connect(transport);
|
||||
|
||||
// Force flush stdout
|
||||
if (process.stdout.isTTY === false) {
|
||||
process.stdout.write('', () => {}); // Force flush
|
||||
}
|
||||
|
||||
logger.info('n8n Documentation MCP Server running on stdio transport');
|
||||
|
||||
// Keep process alive
|
||||
process.stdin.resume();
|
||||
}
|
||||
```
|
||||
|
||||
#### 1.2 Add TTY Support to Docker
|
||||
|
||||
**File:** `Dockerfile`
|
||||
|
||||
Add environment variable to detect Docker:
|
||||
|
||||
```dockerfile
|
||||
ENV IS_DOCKER=true
|
||||
ENV NODE_OPTIONS="--max-old-space-size=2048"
|
||||
```
|
||||
|
||||
**File:** Update Docker command in README
|
||||
|
||||
```json
|
||||
{
|
||||
"mcpServers": {
|
||||
"n8n-mcp": {
|
||||
"command": "docker",
|
||||
"args": [
|
||||
"run",
|
||||
"--rm",
|
||||
"-i",
|
||||
"-t", // Add TTY allocation
|
||||
"--init", // Proper signal handling
|
||||
"-e", "MCP_MODE=stdio",
|
||||
"-e", "LOG_LEVEL=error",
|
||||
"-e", "DISABLE_CONSOLE_OUTPUT=true",
|
||||
"ghcr.io/czlonkowski/n8n-mcp:latest"
|
||||
]
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
### Phase 2: Robust Fixes (Medium Priority)
|
||||
|
||||
#### 2.1 Implement Explicit Initialize Handler
|
||||
|
||||
**File:** `src/mcp/server-update.ts`
|
||||
|
||||
Add explicit initialize handler to ensure response:
|
||||
|
||||
```typescript
|
||||
import {
|
||||
InitializeRequestSchema,
|
||||
InitializeResult
|
||||
} from '@modelcontextprotocol/sdk/types.js';
|
||||
|
||||
private setupHandlers(): void {
|
||||
// Add explicit initialize handler
|
||||
this.server.setRequestHandler(InitializeRequestSchema, async (request) => {
|
||||
logger.debug('Handling initialize request', request);
|
||||
|
||||
const result: InitializeResult = {
|
||||
protocolVersion: "2024-11-05",
|
||||
capabilities: {
|
||||
tools: {}
|
||||
},
|
||||
serverInfo: {
|
||||
name: "n8n-documentation-mcp",
|
||||
version: "1.0.0"
|
||||
}
|
||||
};
|
||||
|
||||
// Force immediate flush
|
||||
if (process.stdout.isTTY === false) {
|
||||
process.stdout.write('', () => {});
|
||||
}
|
||||
|
||||
return result;
|
||||
});
|
||||
|
||||
// ... existing handlers
|
||||
}
|
||||
```
|
||||
|
||||
#### 2.2 Add Docker-Specific Stdio Handling
|
||||
|
||||
**File:** Create `src/utils/docker-stdio.ts`
|
||||
|
||||
```typescript
|
||||
export class DockerStdioTransport extends StdioServerTransport {
|
||||
constructor() {
|
||||
super();
|
||||
|
||||
// Disable buffering for Docker
|
||||
if (process.env.IS_DOCKER === 'true') {
|
||||
process.stdout.setDefaultEncoding('utf8');
|
||||
if (process.stdout._handle && process.stdout._handle.setBlocking) {
|
||||
process.stdout._handle.setBlocking(true);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
protected async writeMessage(message: string): Promise<void> {
|
||||
await super.writeMessage(message);
|
||||
|
||||
// Force flush in Docker
|
||||
if (process.env.IS_DOCKER === 'true') {
|
||||
process.stdout.write('', () => {});
|
||||
}
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
### Phase 3: Alternative Approaches (Low Priority)
|
||||
|
||||
#### 3.1 Use Wrapper Script
|
||||
|
||||
Create a Node.js wrapper that ensures proper buffering:
|
||||
|
||||
**File:** `docker-entrypoint.js`
|
||||
|
||||
```javascript
|
||||
#!/usr/bin/env node
|
||||
|
||||
// Disable all buffering
|
||||
process.stdout._handle?.setBlocking?.(true);
|
||||
process.stdin.setRawMode?.(false);
|
||||
|
||||
// Import and run the actual server
|
||||
require('./dist/mcp/index.js');
|
||||
```
|
||||
|
||||
#### 3.2 Switch to HTTP Transport for Docker
|
||||
|
||||
Consider using HTTP transport instead of stdio for Docker deployments, as it doesn't have buffering issues.
|
||||
|
||||
## Testing Plan
|
||||
|
||||
1. **Local Testing:**
|
||||
```bash
|
||||
# Test with Docker TTY
|
||||
docker run -it --rm ghcr.io/czlonkowski/n8n-mcp:latest
|
||||
|
||||
# Test initialize response
|
||||
echo '{"jsonrpc":"2.0","method":"initialize","params":{"protocolVersion":"2024-11-05"},"id":1}' | \
|
||||
docker run -i --rm ghcr.io/czlonkowski/n8n-mcp:latest
|
||||
```
|
||||
|
||||
2. **Claude Desktop Testing:**
|
||||
- Apply fixes incrementally
|
||||
- Test with each configuration change
|
||||
- Monitor Claude Desktop logs
|
||||
|
||||
3. **Debug Output:**
|
||||
Add temporary debug logging to stderr:
|
||||
```typescript
|
||||
console.error('DEBUG: Received initialize');
|
||||
console.error('DEBUG: Sending response');
|
||||
```
|
||||
|
||||
## Implementation Priority
|
||||
|
||||
1. **Immediate:** Add `-t` flag to Docker command (no code changes)
|
||||
2. **High:** Force stdout flushing in server code
|
||||
3. **Medium:** Add explicit initialize handler
|
||||
4. **Low:** Create Docker-specific transport class
|
||||
|
||||
## Success Criteria
|
||||
|
||||
- Claude Desktop connects without timeout
|
||||
- No "Shutting down..." message in JSON stream
|
||||
- Tools are accessible after connection
|
||||
- Connection remains stable
|
||||
|
||||
## References
|
||||
|
||||
- [Docker stdout buffering issue](https://github.com/moby/moby/issues/1385)
|
||||
- [MCP TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk)
|
||||
- [Python unbuffered mode in Docker](https://stackoverflow.com/questions/39486327/stdout-being-buffered-in-docker-container)
|
||||
- [MCP initialization timeout issues](https://github.com/modelcontextprotocol/servers/issues/57)
|
||||
@@ -1,132 +0,0 @@
|
||||
# HTTP Server Final Fix Documentation
|
||||
|
||||
## Problem Summary
|
||||
|
||||
The n8n-MCP HTTP server experienced two critical issues:
|
||||
|
||||
1. **"stream is not readable" error** - Caused by Express.json() middleware consuming the request stream
|
||||
2. **"Server not initialized" error** - Caused by StreamableHTTPServerTransport's internal state management
|
||||
|
||||
## Solution Overview
|
||||
|
||||
We implemented a two-phase fix:
|
||||
|
||||
### Phase 1: Stream Preservation (v2.3.2)
|
||||
- Removed global `express.json()` middleware
|
||||
- Allowed StreamableHTTPServerTransport to read raw request stream
|
||||
- This fixed the "stream is not readable" error but revealed the initialization issue
|
||||
|
||||
### Phase 2: Direct JSON-RPC Implementation
|
||||
- Created `http-server-fixed.ts` that bypasses StreamableHTTPServerTransport
|
||||
- Implements JSON-RPC protocol directly
|
||||
- Handles MCP methods: initialize, tools/list, tools/call
|
||||
- Maintains full protocol compatibility
|
||||
|
||||
## Implementation Details
|
||||
|
||||
### The Fixed Server (`http-server-fixed.ts`)
|
||||
|
||||
```javascript
|
||||
// Instead of using StreamableHTTPServerTransport
|
||||
const transport = new StreamableHTTPServerTransport({...});
|
||||
|
||||
// We handle JSON-RPC directly
|
||||
req.on('data', chunk => body += chunk);
|
||||
req.on('end', () => {
|
||||
const jsonRpcRequest = JSON.parse(body);
|
||||
// Handle request based on method
|
||||
});
|
||||
```
|
||||
|
||||
### Key Features:
|
||||
1. **No body parsing middleware** - Preserves raw stream
|
||||
2. **Direct JSON-RPC handling** - Avoids transport initialization issues
|
||||
3. **Persistent MCP server** - Single instance handles all requests
|
||||
4. **Manual method routing** - Implements initialize, tools/list, tools/call
|
||||
|
||||
### Supported Methods:
|
||||
- `initialize` - Returns server capabilities
|
||||
- `tools/list` - Returns available tools
|
||||
- `tools/call` - Executes specific tools
|
||||
|
||||
## Usage
|
||||
|
||||
### Environment Variables:
|
||||
- `MCP_MODE=http` - Enable HTTP mode
|
||||
- `USE_FIXED_HTTP=true` - Use the fixed implementation
|
||||
- `AUTH_TOKEN` - Authentication token (32+ chars recommended)
|
||||
|
||||
### Starting the Server:
|
||||
```bash
|
||||
# Local development
|
||||
MCP_MODE=http USE_FIXED_HTTP=true AUTH_TOKEN=your-secure-token npm start
|
||||
|
||||
# Docker
|
||||
docker run -d \
|
||||
-e MCP_MODE=http \
|
||||
-e USE_FIXED_HTTP=true \
|
||||
-e AUTH_TOKEN=your-secure-token \
|
||||
-p 3000:3000 \
|
||||
ghcr.io/czlonkowski/n8n-mcp:latest
|
||||
```
|
||||
|
||||
### Testing:
|
||||
```bash
|
||||
# Initialize
|
||||
curl -X POST http://localhost:3000/mcp \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer your-secure-token" \
|
||||
-d '{"jsonrpc":"2.0","method":"initialize","params":{},"id":1}'
|
||||
|
||||
# List tools
|
||||
curl -X POST http://localhost:3000/mcp \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer your-secure-token" \
|
||||
-d '{"jsonrpc":"2.0","method":"tools/list","id":2}'
|
||||
|
||||
# Call tool
|
||||
curl -X POST http://localhost:3000/mcp \
|
||||
-H "Content-Type: application/json" \
|
||||
-H "Authorization: Bearer your-secure-token" \
|
||||
-d '{"jsonrpc":"2.0","method":"tools/call","params":{"name":"get_node_info","arguments":{"nodeType":"httpRequest"}},"id":3}'
|
||||
```
|
||||
|
||||
## Technical Details
|
||||
|
||||
### Why StreamableHTTPServerTransport Failed
|
||||
|
||||
1. **Stateful Design**: The transport expects to maintain session state
|
||||
2. **Initialization Sequence**: Requires specific handshake before accepting requests
|
||||
3. **Stream Consumption**: Conflicts with Express middleware patterns
|
||||
4. **Version Issues**: Despite fixes in v1.10.1+, issues persist with stateless usage
|
||||
|
||||
### Why Direct Implementation Works
|
||||
|
||||
1. **No Middleware Conflicts**: We control the entire request lifecycle
|
||||
2. **No State Requirements**: Each request is handled independently
|
||||
3. **Protocol Compliance**: Still implements standard JSON-RPC 2.0
|
||||
4. **Simplicity**: Fewer moving parts mean fewer failure points
|
||||
|
||||
## Performance Characteristics
|
||||
|
||||
- **Memory Usage**: ~10-20MB base, grows with database queries
|
||||
- **Response Time**: <50ms for most operations
|
||||
- **Concurrent Requests**: Handles multiple requests without session conflicts
|
||||
- **Database Access**: Single persistent connection, no connection overhead
|
||||
|
||||
## Future Considerations
|
||||
|
||||
1. **Streaming Support**: Current implementation doesn't support SSE streaming
|
||||
2. **Session Management**: Could add optional session tracking if needed
|
||||
3. **Protocol Extensions**: Easy to add new JSON-RPC methods
|
||||
4. **Migration Path**: Can switch back to StreamableHTTPServerTransport when fixed
|
||||
|
||||
## Conclusion
|
||||
|
||||
The fixed implementation provides a stable, production-ready HTTP server for n8n-MCP that:
|
||||
- Works reliably without stream errors
|
||||
- Maintains MCP protocol compatibility
|
||||
- Simplifies debugging and maintenance
|
||||
- Provides better performance characteristics
|
||||
|
||||
This solution demonstrates that sometimes bypassing problematic libraries and implementing core functionality directly is the most pragmatic approach.
|
||||
@@ -1,32 +0,0 @@
|
||||
# Docker Fix for v2.6.2 - Missing Runtime Dependencies
|
||||
|
||||
## Issue
|
||||
After v2.6.0, the Docker image started failing with:
|
||||
```
|
||||
Error: Cannot find module 'axios'
|
||||
```
|
||||
|
||||
## Root Cause
|
||||
In v2.6.0, we added n8n management tools that require:
|
||||
- `axios` - For HTTP API calls to n8n instances
|
||||
- `zod` - For workflow validation schemas
|
||||
|
||||
However, our ultra-optimized Docker image uses `package.runtime.json` which didn't include these new dependencies.
|
||||
|
||||
## Fix
|
||||
Updated `package.runtime.json` to include:
|
||||
```json
|
||||
"axios": "^1.10.0",
|
||||
"zod": "^3.25.32"
|
||||
```
|
||||
|
||||
## Impact
|
||||
- Docker image size increases slightly (~5MB) but remains much smaller than full n8n dependencies
|
||||
- No changes needed to Dockerfile itself - just the runtime package list
|
||||
- Users need to pull the latest Docker image after rebuild
|
||||
|
||||
## Prevention
|
||||
When adding new features that require additional npm packages, always check:
|
||||
1. Is the package needed at runtime?
|
||||
2. If yes, add it to `package.runtime.json` for Docker builds
|
||||
3. Test the Docker image before release
|
||||
@@ -1,135 +0,0 @@
|
||||
# Phase 2 Improvements - v2.4.2
|
||||
|
||||
## 🎯 Overview
|
||||
|
||||
Following the successful implementation of operation-aware validation, Phase 2 adds professional-grade features that make the validation system even more powerful and flexible.
|
||||
|
||||
## ✅ Implemented Features
|
||||
|
||||
### 1. **Validation Profiles** 🎨
|
||||
|
||||
Different validation levels for different use cases:
|
||||
|
||||
```typescript
|
||||
validate_node_operation({
|
||||
nodeType: "nodes-base.slack",
|
||||
config: { ... },
|
||||
profile: "minimal" // or "runtime", "ai-friendly", "strict"
|
||||
})
|
||||
```
|
||||
|
||||
#### Available Profiles:
|
||||
|
||||
| Profile | Purpose | What it checks |
|
||||
|---------|---------|----------------|
|
||||
| **minimal** | Quick check | Only missing required fields |
|
||||
| **runtime** | Pre-execution | Critical errors + security warnings |
|
||||
| **ai-friendly** | Balanced (default) | Errors + helpful warnings |
|
||||
| **strict** | Code review | Everything + best practices |
|
||||
|
||||
### 2. **New Node Validators** 🔧
|
||||
|
||||
Added comprehensive validators for commonly used nodes:
|
||||
|
||||
#### **Webhook Validator**
|
||||
- Path format validation (no spaces, special chars)
|
||||
- Response mode checks
|
||||
- HTTP method validation
|
||||
- Authentication warnings
|
||||
|
||||
#### **PostgreSQL Validator**
|
||||
- SQL injection detection
|
||||
- DELETE/UPDATE without WHERE warnings
|
||||
- Operation-specific validation (insert, update, delete, execute)
|
||||
- Query safety checks
|
||||
|
||||
#### **MySQL Validator**
|
||||
- Similar to PostgreSQL
|
||||
- MySQL-specific syntax checks
|
||||
- Timezone configuration suggestions
|
||||
|
||||
### 3. **validate_node_minimal Tool** ⚡
|
||||
|
||||
Lightning-fast validation for just required fields:
|
||||
|
||||
```json
|
||||
{
|
||||
"nodeType": "nodes-base.slack",
|
||||
"displayName": "Slack",
|
||||
"valid": false,
|
||||
"missingRequiredFields": ["Channel"]
|
||||
}
|
||||
```
|
||||
|
||||
- No warnings
|
||||
- No suggestions
|
||||
- No examples
|
||||
- Just missing required fields
|
||||
- Perfect for quick checks
|
||||
|
||||
### 4. **SQL Safety Features** 🛡️
|
||||
|
||||
Comprehensive SQL query validation:
|
||||
- Detects template expressions that could be vulnerable
|
||||
- Warns about DELETE/UPDATE without WHERE
|
||||
- Catches dangerous operations (DROP, TRUNCATE)
|
||||
- Suggests parameterized queries
|
||||
- Database-specific checks (PostgreSQL $$ quotes, MySQL backticks)
|
||||
|
||||
## 📊 Impact
|
||||
|
||||
### Before Phase 2:
|
||||
- Single validation mode
|
||||
- Limited node coverage (4 nodes)
|
||||
- No SQL safety checks
|
||||
- Fixed validation behavior
|
||||
|
||||
### After Phase 2:
|
||||
- 4 validation profiles for different needs
|
||||
- 7+ nodes with specific validators
|
||||
- Comprehensive SQL injection prevention
|
||||
- Flexible validation based on use case
|
||||
- Ultra-fast minimal validation option
|
||||
|
||||
## 🚀 Usage Examples
|
||||
|
||||
### Using Validation Profiles:
|
||||
```javascript
|
||||
// Quick check - just required fields
|
||||
validate_node_minimal({
|
||||
nodeType: "nodes-base.webhook",
|
||||
config: { responseMode: "lastNode" }
|
||||
})
|
||||
// Result: Missing required field "path"
|
||||
|
||||
// Pre-execution validation
|
||||
validate_node_operation({
|
||||
nodeType: "nodes-base.postgres",
|
||||
config: {
|
||||
operation: "execute",
|
||||
query: "DELETE FROM users WHERE id = ${userId}"
|
||||
},
|
||||
profile: "runtime"
|
||||
})
|
||||
// Result: SQL injection warning
|
||||
|
||||
// Strict validation for code review
|
||||
validate_node_operation({
|
||||
nodeType: "nodes-base.slack",
|
||||
config: { /* valid config */ },
|
||||
profile: "strict"
|
||||
})
|
||||
// Result: Suggestions for best practices
|
||||
```
|
||||
|
||||
## 🎉 Summary
|
||||
|
||||
Phase 2 transforms the validation system from a simple checker into a comprehensive validation framework:
|
||||
|
||||
1. **Flexibility** - Choose validation level based on your needs
|
||||
2. **Safety** - SQL injection detection and prevention
|
||||
3. **Speed** - Minimal validation for quick checks
|
||||
4. **Coverage** - More nodes with specific validation logic
|
||||
5. **Intelligence** - Context-aware suggestions and warnings
|
||||
|
||||
The validation system now provides professional-grade safety and flexibility while maintaining the simplicity that makes it useful for AI agents.
|
||||
Reference in New Issue
Block a user