What happened
AI pentesting vendor Equixly's August 2026 release (published Aug 24, 2026) added Attack Trace (showing the agent's reasoning path to a finding, distinguishing agent-reasoning steps from algorithmic checks) and made its always-on AI Red Teaming of detected LLM endpoints visible and configurable (general behavior, guardrails, system prompt) in the UI, alongside scan-workflow and MCP-server usability improvements.
Why it matters
Improves explainability and control for agentic pentesting findings and formalizes previously implicit automatic LLM red-teaming, aiding triage and handoff for security teams evaluating AI-driven pentest platforms.
Applicability
Existing Equixly customers and teams evaluating agentic AI pentesting/red-teaming platforms.