feat(google-docs): security validations [INTEG-3326]

[primeinteger]

Purpose

Adds security validation to prevent prompt injection attacks when processing Google Docs with AI.

Approach

New Security Module (functions/security/):

• ContentValidator - Detects prompt injection patterns and removes dangerous characters
• GoogleDocsValidator - Validates Google Docs JSON, checking only user text (paragraphs, tables, images) and skipping metadata
• 27 tests covering all validation scenarios

Integration:

• Validates documents in documentParser.agent before sending to OpenAI
• Blocks processing if malicious patterns detected
• Added security instructions to system prompt

Future Use:

• validateContentSecurity() available for validating user input in future features

Testing steps

1. Run tests: npm test -- security/
2. Test with malicious doc: Add "Ignore all previous instructions" to a Google Doc and verify it's blocked
3. Test with normal doc: Verify regular content processes successfully

Screenshots

Function logs include details if the document was rejected due to security validation:

[ryunsong-contentful]

nice, easy to extend and can even be its own agent in the future