Context Search Tool

The Context Search tool (noodlbox_query_with_context) performs intelligent full-text searches that understand your current task and goals. Unlike simple text matching, this tool ranks results based on relevance to what you're trying to accomplish.

Parameters

Parameter	Type	Required	Description
`repository`	string	Yes	Repository name to search
`q`	string	Yes	Natural language search query
`task_context`	string	Yes	What task you're working on
`current_goal`	string	Yes	Your current objective
`search_intention`	string	Yes	Why you're searching
`include_content`	boolean	No	Include full symbol content (default: false)
`limit`	number	No	Max processes to return (1-100, default: 10)
`max_symbols`	number	No	Symbols per process (1-100, default: 20)

Basic Usage

Find Authentication Code

{
  "repository": "my-app",
  "q": "user authentication login",
  "task_context": "Adding OAuth support",
  "current_goal": "Understand existing auth flow",
  "search_intention": "Find auth-related functions to integrate with"
}

Locate Error Handling

{
  "repository": "my-app",
  "q": "error handling exceptions",
  "task_context": "Debugging production error",
  "current_goal": "Find where errors are caught",
  "search_intention": "Trace error propagation path"
}

Context Parameters Explained

task_context

Describe what you're currently working on. This helps the search understand which results are most relevant.

Examples:

"Adding new payment integration"
"Fixing memory leak in data processing"
"Refactoring authentication module"

current_goal

Specify your immediate objective. This focuses the search on results that help achieve this goal.

Examples:

"Understand how payments are processed"
"Find all database connection points"
"Locate test setup code"

search_intention

Explain why you're searching. This helps rank results by how useful they'll be.

Examples:

"Need to add similar functionality"
"Looking for examples of this pattern"
"Finding code to refactor"
"Understanding system behavior"

Response Format

Results are organized hierarchically: Processes -> Symbols

interface ContextSearchResponse {
  processes: ProcessResult[];
  total_processes: number;
  truncated: boolean;
}

interface ProcessResult {
  process_id: string;
  process_label: string;
  process_type: 'intra_community' | 'cross_community';
  relevance_score: number;
  symbols: SymbolResult[];
}

interface SymbolResult {
  symbol_name: string;
  symbol_type: string;  // "Function", "Class", "Variable", etc.
  file_path: string;
  line_number?: number;
  content?: string;  // If include_content is true
}

Example Response

{
  "processes": [
    {
      "process_id": "abc123",
      "process_label": "UserAuthentication",
      "process_type": "intra_community",
      "relevance_score": 0.95,
      "symbols": [
        {
          "symbol_name": "authenticateUser",
          "symbol_type": "Function",
          "file_path": "src/auth/authenticate.ts",
          "line_number": 45
        },
        {
          "symbol_name": "validateToken",
          "symbol_type": "Function",
          "file_path": "src/auth/token.ts",
          "line_number": 23
        }
      ]
    }
  ],
  "total_processes": 1,
  "truncated": false
}

Advanced Features

Process-Grouped Results

Results are automatically grouped by process, making it easier to understand how code fits together:

{
  "q": "database query",
  "task_context": "Optimizing database performance",
  "current_goal": "Find slow queries",
  "search_intention": "Identify queries to optimize"
}

This returns processes like "DatabaseAccess", "UserRepository", etc., each containing related symbols.

Pagination

For large result sets, use pagination:

{
  "q": "api endpoint",
  "limit": 5,  // Return 5 processes
  "max_symbols": 10  // Up to 10 symbols per process
}

Including Symbol Content

Get full symbol definitions by setting include_content:

{
  "q": "payment processing",
  "include_content": true
}

Including content significantly increases response size. Use sparingly for targeted searches.

Search Strategies

Strategy 1: Broad Discovery

Start broad to understand what's available:

{
  "q": "payment",
  "task_context": "Adding payment features",
  "current_goal": "Explore payment-related code",
  "search_intention": "Get overview of payment system",
  "limit": 20
}

Strategy 2: Focused Investigation

Narrow down to specific functionality:

{
  "q": "stripe webhook handler",
  "task_context": "Fixing webhook processing",
  "current_goal": "Find webhook handler code",
  "search_intention": "Debug webhook failures",
  "limit": 5,
  "include_content": true
}

Strategy 3: Pattern Matching

Find similar implementations:

{
  "q": "react component form validation",
  "task_context": "Creating new form",
  "current_goal": "Find validation examples",
  "search_intention": "Reuse existing validation patterns"
}

Best Practices

1. Provide Rich Context

Good:

{
  "q": "user profile",
  "task_context": "Implementing profile edit feature",
  "current_goal": "Understand existing profile structure",
  "search_intention": "Find profile models and update logic"
}

Less effective:

{
  "q": "user profile",
  "task_context": "Working on code",
  "current_goal": "Find stuff",
  "search_intention": "Looking around"
}

2. Use Natural Language

The search understands natural language queries:

"how to handle user authentication"
"error logging and monitoring"
"database connection pooling"

3. Iterate Your Search

Refine searches based on results:

Start broad: "authentication"
Focus: "JWT token validation"
Specific: "token expiry handling"

4. Adjust Result Limits

Balance comprehensiveness with performance:

Quick overview: limit: 5
Thorough investigation: limit: 20
Exhaustive search: limit: 100

Troubleshooting

Too Many Results

Make query more specific
Reduce limit and max_symbols
Add more detail to context parameters

Not Finding Expected Code

Try different search terms
Broaden context parameters
Increase limit parameter
Try searching for file names or specific function names

Results Not Relevant

Refine task_context to be more specific
Adjust current_goal to better describe what you need
Make search_intention more explicit

Performance Tips

Start small: Use low limits initially, increase as needed
Skip content: Don't use include_content unless necessary
Be specific: More specific queries return faster
Use pagination: Fetch results in batches for large searches

Cypher Query Tool - For more precise graph queries
Codebase Map Resource - Understand code structure
Code Exploration - Discovery workflows

On this page