Add sanitizeError() and sanitizeMessage() functions that strip sensitive
information from error messages before they are stored in scraper_runs
history. This prevents accidental exposure of infrastructure details in
the database.
Sanitization covers:
- Unix and Windows file paths (e.g., /home/deploy/app/..., C:\Users\...)
- MongoDB connection strings (mongodb:// and mongodb+srv://)
- Credential patterns (API_KEY=, password=, secret=, token=)
- Stack trace file path references
The error type/name (e.g., MongoServerError, TypeError) and general
debugging description are preserved to maintain diagnostic usefulness.
Applied in runScrape() catch block before calling recordScraperRun(),
ensuring only sanitized messages reach the database.
Added 18 new tests covering all sanitization categories: file paths,
connection strings, credentials, error type preservation, description
preservation, stack trace removal, and integration with recordScraperRun.
