Turn specs into evals for any agent with ASSERT

Microsoft Security Blog published a set of posts (June 2026) focused on AI/agent security, CI/CD and software supply-chain threats, and investigative tooling. Key items include: ASSERT, an open-source framework to turn natural-language specs into executable evaluations for agents; a telemetry-driven playbook for reconstructing AI activity in Microsoft 365 Copilot and Azure AI; analysis of threat actors using AI branding as social-engineering bait; discovery of a prompt-injection pathway in the Claude Code GitHub Action that could expose workflow secrets and the related responsible disclosure/m