msbaek/databricks-academy

Databricks Academy

Overview

This skill enables access to login-protected content on Databricks Academy (https://customer-academy.databricks.com/). It handles authentication, searches for courses and tutorials, and retrieves learning materials to answer user questions about Databricks topics.

When to Use This Skill

Use this skill when:

• Users ask about Databricks courses, learning paths, or certifications
• Users need to find tutorials or documentation for Databricks features (Delta Lake, Unity Catalog, etc.)
• Users want to explore hands-on labs or practical exercises
• Users inquire about specific topics like "SQL Analytics", "MLflow", "Data Pipelines"
• Any query that requires accessing content from customer-academy.databricks.com

Workflow

Step 1: Collect Login Credentials

Before accessing Databricks Academy, collect authentication credentials from the user:

1. Use AskUserQuestion to request email and password
2. Inform the user: "These credentials are used only for this session and will not be stored"
3. Handle sensitive information securely - never log or store passwords

Security Note: Credentials are kept in memory only for the duration of the browser session and discarded when the session ends.

Step 2: Authenticate to Databricks Academy

Perform browser automation to log in:

1. Navigate to https://customer-academy.databricks.com/ using browser_navigate
2. Capture current page state with browser_snapshot
3. Locate login form elements (email and password fields)
4. Enter email using browser_type with the email field reference
5. Enter password using browser_type with the password field reference
6. Submit login form using browser_click on the login button
7. Wait for navigation completion using browser_wait_for
8. Verify successful login with browser_snapshot

Error Handling: If login fails:

• Inform the user: "Login failed. Please verify your credentials and try again."
• Check if the account requires additional verification (2FA, SSO redirect)
• Suggest the user verify access by logging in directly to the website first

Step 3: Search for Content

After successful authentication, search for the requested content:

1. Use browser_snapshot to identify the search interface
2. Locate the search input field
3. Enter the user's query using browser_type
4. Execute search using browser_click
5. Wait for results to load using browser_wait_for
6. Capture search results with browser_snapshot

Search Optimization:

• Use specific English keywords for better results (see references/common_queries.md)
• If initial query returns no results, try related terms:
  • "Delta Lake" → "Introduction to Delta Lake", "Delta Lake Tutorial"
  • "SQL" → "SQL Analytics", "Databricks SQL", "SQL Fundamentals"
• Search for course codes if known (e.g., "DE 101", "ML 201")

Step 4: Extract and Present Results

Process search results and present them to the user:

1. Extract relevant information from browser_snapshot:

   • Course/tutorial titles
   • Descriptions or summaries
   • URLs (relative or absolute)
   • Content type (course, lab, documentation, certification prep)

2. Format results in structured markdown:

   ## Search Results for "[query]"
   
   ### 1. [Course/Tutorial Title]
   - **Type**: [Course/Lab/Documentation]
   - **Link**: [URL]
   - **Description**: [Brief description]
   - **Key Topics**: [Main concepts covered]
   
   ### 2. [Next Result]
   ...
   
   ### Recommendations
   - [Suggested follow-up courses or related materials]
   

3. For course detail requests, navigate to the specific course page:

   • Use browser_click to open the course
   • Extract syllabus, modules, or learning objectives
   • Summarize the content structure

Step 5: Handle Follow-up Queries

During the session, maintain the browser context to handle additional requests:

• Keep the browser session alive for follow-up questions
• Navigate to different pages as needed using browser_navigate or browser_click
• Extract specific information from course pages or modules
• Close the browser only when the user ends the session or no longer needs access

Session Management: After completing all queries, close the browser using browser_close to ensure credentials are discarded.

Error Handling

Login Failures

Symptoms: Unable to authenticate, wrong credentials, or unexpected redirects

Actions:

• Verify credentials with the user
• Check if the website structure has changed (different login form elements)
• Inform the user if SSO or 2FA is blocking automated login
• Suggest manual verification on the website

Page Load Issues

Symptoms: Timeouts, incomplete page loads, dynamic content not appearing

Actions:

• Increase wait time using browser_wait_for with appropriate text or time parameter
• Use browser_snapshot to verify current page state
• Retry navigation if needed
• Inform the user if the website is experiencing issues

Search Returns No Results

Symptoms: Empty search results or "no content found" messages

Actions:

• Try alternative search terms (see references/common_queries.md)
• Suggest broader or narrower queries
• Recommend specific courses or learning paths related to the topic
• Verify the topic is covered on Databricks Academy

Browser Session Errors

Symptoms: Browser crashes, lost connection, or unexpected closures

Actions:

• Restart the browser and re-authenticate
• Inform the user about the interruption
• Resume from the last known state

Tools Used

This skill relies on the following MCP tools:

• mcp__playwright__browser_navigate - Navigate to URLs
• mcp__playwright__browser_snapshot - Capture page structure
• mcp__playwright__browser_type - Enter text in form fields
• mcp__playwright__browser_click - Click buttons and links
• mcp__playwright__browser_wait_for - Wait for page elements or time delays
• mcp__playwright__browser_close - Close browser and clear session
• AskUserQuestion - Collect user input securely

References

Common Queries

See references/common_queries.md for:

• Popular Databricks Academy topics (Data Engineering, ML, Analytics)
• Effective search keywords
• Course structure and content types

Use this reference to optimize searches and suggest related content to users.

Limitations

1. Authentication Required: All content requires a valid Databricks Academy account
2. Access Restrictions: Content availability depends on user account permissions and subscriptions
3. Dynamic Content: Page structure may change, requiring adjustments to element selectors
4. Session Duration: Browser sessions are temporary and must be re-authenticated for new conversations
5. Rate Limiting: Excessive requests may trigger website rate limiting or security measures

Best Practices

• Always verify successful login before proceeding with searches
• Use browser_snapshot frequently to adapt to page structure
• Provide clear, structured output to users
• Suggest related courses and learning paths proactively
• Close browser sessions promptly after completing requests
• Never log or store user credentials