---
name: mcp-endpoint-explorer
description: Explore and find API endpoints using MCP servers. Lists collections, searches for endpoints, extracts endpoint details. Use when searching for API endpoints, exploring Postman collections, finding specific routes, or discovering available endpoints in MCP servers.
---

# MCP Endpoint Explorer

Explores and discovers API endpoints through MCP servers, particularly Postman collections.

## Workflow

### 1. Discover Available MCP Servers

Check available MCP servers and their tools:

- List MCP server directories in `/root/.cursor/projects/root-projects-Tangerina/mcps/`
- Review tool descriptors to understand available capabilities

### 2. List Collections/Workspaces

For Postman MCP server:

1. Call `getWorkspaces` to list available workspaces
2. Use workspace ID to call `getCollections` with workspace parameter
3. Note collection IDs for detailed exploration

### 3. Get Collection Details

Use `getCollection` with:

- `collectionId`: Format `<OWNER_ID>-<UUID>`
- `model`: `minimal` for structure, `full` for complete details

### 4. Search for Endpoints

Search strategies:

- **By name**: Search collection JSON for request names containing keywords
- **By path**: Search URL paths for specific route patterns
- **By method**: Filter by HTTP method (GET, POST, etc.)
- **By query params**: Search for specific query parameters

### 5. Extract Endpoint Information

For each found endpoint, extract:

- Request name
- HTTP method
- Full URL pattern (with variables)
- Path segments
- Query parameters (name, value, required/optional)
- Authentication method
- Request body structure (if applicable)

## Common Patterns

### Finding Endpoints by Domain

```bash
# Search collection file for domain keywords
grep -i "keyword" collection.json | grep -E "(name|url|method)"
```

### Extracting Complete Endpoint Details

Use Python to parse JSON and extract structured endpoint information:

```python
import json

def find_endpoints(collection_data, search_term):
    # Recursively search through collection items
    # Return matching requests with full details
```

## Utility Script

Use `scripts/retrieve-endpoints.sh` to:

- List all collections
- Export collection data
- Search for specific endpoints
- Generate endpoint documentation

Run: `bash scripts/retrieve-endpoints.sh [search-term]`

## Output Format

When presenting found endpoints:

**Endpoint Name:** [Request Name]
**Method:** [HTTP Method]
**Path:** [Endpoint Path]
**Full URL:** [Complete URL with variables]
**Query Parameters:**

- `param1`: description (required/optional)
- `param2`: description (required/optional)
  **Authentication:** [Auth type and token variable]

## Best Practices

1. **Start broad**: List collections before searching
2. **Use minimal model first**: Get structure before full details
3. **Search systematically**: Use multiple search strategies
4. **Extract complete info**: Include all parameters and auth details
5. **Document variables**: Note Postman variables used in URLs