Content Security Policy (CSP) and Cross-Origin Resource Sharing (CORS) improvements

CVE remediation release

Enhancement:

• Introduced a feature flag for the PII rule that enables the administrators to toggle between the standard and the "strict" mode

Bug fix:

• Fixed the email address PII detection issue that was not catching addresses with a certain format

Vulnerability Fix:

• Patched pillow vulnerability CVE-2025-48379

Enhancements:

• The profanity model is now embedded inside the container image instead of downloaded remotely at server startup
• Prompt injection check now uses a sliding window chunking iterator to iterate over an entire text instead of validating only the edges of the text

Enhancements:

• Eliminated the API key validation latency that was affecting users with a large number of active API keys

Bug Fix

• The UI now correctly displays the inference validation status when the rule validation result is in a "skipped" state

IMPORTANT Note for On-Premises Customers:

There's a new tuning parameter for the toxicity detection feature. Please update the "Shield Toxicity Check Max Token Limit" deployment configuration value from 1,200 to 8,000.

CFT: ShieldToxicityCheckMaxTokenLimit (Shield Toxicity Check Max Token Limit)
Helm Chart: shieldToxicityCheckMaxTokenLimit

Enhancements:

• We’ve made several updates to reduce latency, support longer inputs, and increase performance in our toxicity model:
  • Introduced a new chunking strategy for the harmful request classifier
  • Increased the toxicity classifier’s token limit from 1,200 to 8,000 tokens
  • Upgraded our profanity model resulting in increased performance for profanity detection
• Improved PII Detection Accuracy with New Post-processing Enhancements:
  • Cleaner Entity Recognition: Common non-name terms like “me,” “you,” and “doctor” are now filtered out to reduce false positives in name detection.
  • Smarter Crypto Wallet & Bank Account Validation: Wallet and bank account entities are now validated for structure, expected length, and presence of unexpected characters or common words like “wallet.”
  • Enhanced URL Detection: URLs are now correctly identified even when missing standard prefixes like http://.

IMPORTANT Note for On-Premises Customers:

The Shield server worker count configuration must be set to 2 when upgrading to this version. The previously recommended value was 5. This number can be set higher than 2 to further increase concurrency when AWS EC2s with more GPU memory than the g4dn.2xlarge are used.

• CFT: ShieldServerWorkerCount (Number of Shield Server Workers)
• Helm Chart: shieldWorkers

Enhancements:

• Made significant performance improvements to the PII detection model, resulting in fewer false positives.
  The inference deep dive table now returns up to 50 rows per page.

Vulnerability Fix:

• Patched the protobuf vulnerability CVE-2025-4565

Enhancements:

Improved the prompt injection classifier model, boosting its accuracy and efficiency. The updated model provides enhanced detection capabilities for identifying potential prompt injection attempts. It now prioritizes precision over recall, effectively reducing false positives where legitimate user inputs are mistakenly flagged, while maintaining robust security against genuine attacks.