Release Notes

Stay up to date with the latest features, improvements, and fixes across all Safeguard products.

July 2026

First-party Security Engines

Safeguard now ships a suite of first-party, defensive security engines, all reporting into one unified findings model and each independently admin-toggleable via server-enforced feature flags:

Application Security Testing (SAST & DAST) - static source analysis and defensive, authorized dynamic testing.
Red Team (Defensive Adversary Emulation) - breach-and-attack simulation and purple-teaming using benign ATT&CK detection canaries (no weaponized payloads), attack-path graph with unified risk correlation, safe recon, and an AI/LLM guardrail-validation harness - gated by a signed rules-of-engagement, scope checks, rate/blast caps, human approval for high-impact steps, a global kill-switch, and full audit.
AI-SPM - model-artifact malware scanning (pickle opcode disassembly, torch-zip walker, safetensors/GGUF validation).
DSPM - sensitive-data classification (PII/PHI, Luhn-validated PCI, entropy-gated secrets) with an enforced redaction mandate.
Runtime Protection (CWPP & CNAPP) - ATT&CK-mapped runtime threat rules and cloud-native risk correlation. The eBPF runtime collector is a Linux follow-on, rolling out.
Package Firewall - inline install-time allow/warn/block for dependencies, with audit and quarantine modes.
AutoTriage - cross-scanner deduplication and noise reduction with a measured reduction metric; never suppresses malware or secrets findings.

January 2026

v3.0.0 - January 3, 2026

Major Release: Griffin AI General Availability

New Features

Griffin AI GA - Purpose-built LLM for supply chain security now available to all customers
AI Remediate 2.0 - Enhanced automated remediation with multi-language support
MCP Server - Model Context Protocol server for AI integrations
Container AI Analysis - Deep learning-based container image analysis