https://keldura.ai/daily/ai-technology/2026-07-31/openai-and-anthropic-disclosures-sharpen-the-ai-agent-safety-debate Keldura Daily en 2026-07-31T06:00:24.368108Z OpenAI and Anthropic disclosures sharpen the AI agent safety debate https://keldura.ai/daily/ai-technology/2026-07-31/researchers-say-llms-have-a-fundamental-security-flaw Keldura Daily en 2026-07-31T06:00:24.368108Z Researchers say LLMs have a fundamental security flaw