DiggingBeaglehttps://diggingbeagle.com/Evidence-first AI-security researchAnthropic's September report shows agentic cyber operations moving from assistance to throughputhttps://diggingbeagle.com/news/anthropic-september-report-shows-agentic-cyber-operations-moving-from-assistance-to-throughput/https://diggingbeagle.com/news/anthropic-september-report-shows-agentic-cyber-operations-moving-from-assistance-to-throughput/The cases are still built from familiar intrusion primitives, but reconnaissance, exploit development and data processing are increasingly delegated to persistent agent workflows.Thu, 10 Sep 2026 00:00:00 GMTStolen AI credentials are now showing up as attacker infrastructurehttps://diggingbeagle.com/updates/stolen-ai-credentials-are-now-showing-up-as-attacker-infrastructure/https://diggingbeagle.com/updates/stolen-ai-credentials-are-now-showing-up-as-attacker-infrastructure/Anthropic's September report documents resale, victim-funded compute and attribution cover as separate values of compromised AI access.Thu, 10 Sep 2026 00:00:00 GMTCheck Point demonstrates a cross-account task channel inside ChatGPT sandboxeshttps://diggingbeagle.com/news/check-point-demonstrates-cross-account-task-channel-inside-chatgpt-sandboxes/https://diggingbeagle.com/news/check-point-demonstrates-cross-account-task-channel-inside-chatgpt-sandboxes/A shared internal package service became a covert path between code-execution environments from different accounts.Tue, 08 Sep 2026 00:00:00 GMTOpenAI says agent-spam disclosure criteria are still being definedhttps://diggingbeagle.com/updates/openai-says-agent-spam-disclosure-criteria-are-still-being-defined/https://diggingbeagle.com/updates/openai-says-agent-spam-disclosure-criteria-are-still-being-defined/OpenAI acknowledged the public-wiki activity and its Chief Scientist separately called for stronger shared safety bars.Sun, 06 Sep 2026 00:00:00 GMTResearchers reconstruct an OpenAI agent message board on the public webhttps://diggingbeagle.com/news/researchers-reconstruct-openai-agent-message-board-on-the-public-web/https://diggingbeagle.com/news/researchers-reconstruct-openai-agent-message-board-on-the-public-web/The report found roughly 18,000 retained public posts from autonomous agents sharing answers and bypass ideas during a web-research task.Fri, 04 Sep 2026 00:00:00 GMTKASS moves smart-contract security from detection to executable exploit testinghttps://diggingbeagle.com/news/kass-moves-smart-contract-security-from-detection-to-executable-exploit-testing/https://diggingbeagle.com/news/kass-moves-smart-contract-security-from-detection-to-executable-exploit-testing/A July paper reports a multi-agent system that retrieves audit knowledge, generates Foundry exploits and repairs or replans them against execution feedback.Fri, 17 Jul 2026 00:00:00 GMTMCP Python SDK advisories expose task ownership and browser-origin gapshttps://diggingbeagle.com/news/mcp-python-sdk-advisories-expose-task-ownership-and-browser-origin-gaps/https://diggingbeagle.com/news/mcp-python-sdk-advisories-expose-task-ownership-and-browser-origin-gaps/Two narrowly scoped advisories show why agent control planes still need ordinary session authorization and transport security.Thu, 16 Jul 2026 00:00:00 GMTCheck Point turns an AI-generated malware idea into a browser-only ransomware proof of concepthttps://diggingbeagle.com/news/check-point-turns-ai-generated-malware-idea-into-browser-only-ransomware-poc/https://diggingbeagle.com/news/check-point-turns-ai-generated-malware-idea-into-browser-only-ransomware-poc/A controlled Android demonstration used legitimate File System Access permissions to encrypt selected images without a native payload.Wed, 01 Jul 2026 00:00:00 GMT