v1.13.1 v1.7.20
high risk
data-privacy-stack/presidioSep 22, 2026 → Sep 22, 20260 days between releases
What's New
Introduces TypeSafe Jev HITL backend replacing chat model JSON output
Breaking Changes
- Audit mechanism changed from chat model JSON to TypeSafe Jev structured scoring
New Features
- TypeSafe Jev HITL audit backend integration
- Structured pass/block scoring logic
- Custom organization policy overlay support
- OpenAI-compatible reviewer compatibility
Improvements
- Enhanced security with built-in rules blocking destructive operations
- Parallel risk and authorization intent judgment for faster audits
- Structured scoring replaces chat model JSON for more reliability
Upgrade Notes
Update code handling audit results as JSON output is replaced by structured scores. Configure custom policies as overlays if using them. The OpenAI reviewer remains available for migration.
The Two Releases
v1.13.1
v1.13.1
Sep 22, 2026· 30 days ago
A maintenance release featuring a redesigned test report UI, a fix for sequential YAML file handling, and updated documentation for model guidance.
Full release notesv1.7.20
Sep 22, 2026· 29 days ago
Integrates TypeSafe Jev as a Human-in-the-Loop (HITL) audit backend for CyberStrikeAI, replacing chat model JSON output with structured pass/block scores based on local rules.
Full release notes