Editing MCP Tools

​

Overview

​

Accessing Your Tools

1. Navigate to your project in Pylar
2. Look for the “MCP Tools” section in the right sidebar
3. Click on the tool you want to view or edit

​

Tool Components Overview

​

Function Name

• Located at the top
• Can be renamed by clicking and editing
• Should be descriptive and unique

​

Description

• Summary of what the tool does
• Visible to AI agents
• Can be edited inline

​

SQL Query

• The query that retrieves data
• Can be edited in the query editor
• Supports parameter placeholders

​

Parameters

• Input definitions for the tool
• Can add, remove, or modify parameters
• Define types and requirements

​

Tool Call Arguments

• Test values (right side)
• Used for testing
• Can be modified

​

Editing Function Name

1. Click on the function name at the top
2. Edit the name directly
3. Save your changes

​

Editing Description

1. Click on the description field
2. Modify the text
3. Save changes

• Be specific about what data is returned
• Mention any important behaviors (e.g., “supports partial matching”)
• Keep it concise but informative

• ❌ “Gets data” (too vague)
• ✅ “Fetches engagement scores and related data filtered by event type” (clear and specific)

​

Editing SQL Query

​

Accessing the Query Editor

1. Find the SQL Query section in the tool view
2. Click to edit (or use the edit button)
3. Modify the query as needed
4. Save your changes

​

Common Edits

-- Original
SELECT engagement_score FROM table0 WHERE event_type = '{event_type}'

-- Enhanced
SELECT engagement_score FROM table0 
WHERE event_type = '{event_type}' 
  AND date >= '{start_date}'
  AND engagement_score > {min_score}

-- Ascending instead of descending
ORDER BY engagement_score ASC

-- Original
SELECT engagement_score FROM table0

-- Enhanced
SELECT engagement_score, event_id, timestamp, user_id 
FROM table0

-- Add LIMIT for large result sets
SELECT engagement_score FROM table0 
WHERE event_type = '{event_type}'
ORDER BY engagement_score DESC
LIMIT 100

​

Modifying Parameters

​

Adding Parameters

1. Go to the Parameters section
2. Click Add Parameter or similar
3. Define:
   • Parameter name
   • Type (string, number, etc.)
   • Whether it’s required
   • Description
4. Update your SQL query to use the new parameter

-- Query with new parameter
SELECT * FROM table0 
WHERE event_type = '{event_type}'
  AND date >= '{start_date}'
  AND date <= '{end_date}'

​

Removing Parameters

1. Find the parameter in the Parameters section
2. Remove it
3. Update your SQL query to remove the placeholder

​

Modifying Parameter Properties

• Type: string → number, etc.
• Required/Optional: Make parameters required or optional
• Description: Update what the parameter represents

​

Updating Tool Call Arguments

1. Find the Tool Call Arguments section (right side)
2. Modify the test values
3. Use these values when testing the tool

{
  "event_type": "login"  // Change to "purchase" to test different scenarios
}

​

Common Editing Scenarios

​

Scenario 1: Refining an AI-Generated Tool

1. Add more filters: Include additional WHERE conditions
2. Change sort order: Modify ORDER BY clause
3. Add columns: Return more data
4. Improve description: Make it clearer for agents

​

Scenario 2: Fixing a Tool That’s Not Working

1. Check parameter names: Ensure placeholders match parameter names
2. Verify SQL syntax: Test query manually in SQL IDE
3. Check data types: Ensure parameter types match query expectations
4. Review filters: Make sure WHERE conditions are valid

​

Scenario 3: Optimizing Performance

1. Add LIMIT: Restrict result set size
2. Refine filters: Make WHERE conditions more selective
3. Remove unnecessary columns: Select only needed data
4. Add indexes: Ensure source views have proper indexes

​

Viewing JSON Structure

1. Click “JSON view” button
2. View the structured JSON representation
3. Useful for:
   • Understanding complete configuration
   • Debugging issues
   • Copying tool structure

{
  "function_name": "fetch_engagement_scores_by_event_type",
  "description": "Fetches engagement scores and related data filtered by event type",
  "query": "SELECT engagement_score FROM table0 WHERE event_type LIKE '%{event_type}%' \nORDER BY engagement_score DESC",
  "parameters": {
    "type": "object",
    "required": ["event_type"],
    "properties": {
      "event_type": {
        "type": "string",
        "description": "Event type to filter by (supports partial matching)"
      }
    }
  },
  "tool_call": {
    "name": "fetch_engagement_scores_by_event_type",
    "args": {
      "event_type": "login"
    }
  }
}

​

Best Practices

​

Before Editing

• ✅ Test the current tool to understand its behavior
• ✅ Review the SQL query in the SQL IDE
• ✅ Check parameter usage in the query

​

While Editing

• ✅ Make incremental changes
• ✅ Test after each significant change
• ✅ Keep parameter names consistent
• ✅ Update descriptions when behavior changes

​

After Editing

• ✅ Always test the tool after changes
• ✅ Verify parameters work correctly
• ✅ Check that results are as expected
• ✅ Update descriptions if needed

​

Next Steps

• Testing Your Tools - Verify your changes work correctly
• Publishing Tools - Make your tools available to agents