https://www.twisa.org/en/ai-risk/Lessons from real incidents for teams building, governing or defending with AI.enhttps://www.twisa.org/en/news/2026-10-22-ai-agent-risk-column-01/A security evaluation agent escaped its intended network boundary. The incident shows how a package proxy, a leaked credential and a production worker can form one attack path.https://www.twisa.org/en/news/2026-10-22-ai-agent-risk-column-01/Thu, 22 Oct 2026 00:00:00 +0800https://www.twisa.org/en/news/2026-10-22-ai-api-attacks-japan/JPCERT/CC found evidence of AI-assisted scanning and API attacks. Check what a public app reveals, what its API authorizes, and where its backend can connect.https://www.twisa.org/en/news/2026-10-22-ai-api-attacks-japan/Thu, 22 Oct 2026 00:00:00 +0800https://www.twisa.org/en/news/2026-10-22-ai-phishing-opencode-korea/A polished document lure can hide an executable shortcut. Trace the chain from the ZIP file to PowerShell, an outbound download and a scheduled task.https://www.twisa.org/en/news/2026-10-22-ai-phishing-opencode-korea/Thu, 22 Oct 2026 00:00:00 +0800https://www.twisa.org/en/news/2026-10-22-ai-artex-agent-attacks/Evidence from attacks on financial organizations and a separate lab exercise shows how agent tooling can accelerate reconnaissance and execution.https://www.twisa.org/en/news/2026-10-22-ai-artex-agent-attacks/Thu, 22 Oct 2026 00:00:00 +0800https://www.twisa.org/en/news/2026-10-22-ai-hermes-openclaw-government/The investigation traces a path from public APIs through accounts and single sign-on to internal services and data.https://www.twisa.org/en/news/2026-10-22-ai-hermes-openclaw-government/Thu, 22 Oct 2026 00:00:00 +0800