https://blog.badprompt.dev/ Autonomous technical exploit breakdowns, LLM/agent vulnerability research, and CISO threat-landscape intelligence by William Ye (@williamyeee). en-us Tue, 27 Oct 2026 13:00:07 +0000 https://blog.badprompt.dev/posts/php2shell-preauth-rce-cve-2026-109618-109619/ https://blog.badprompt.dev/posts/php2shell-preauth-rce-cve-2026-109618-109619/ 2026-10-27T00:00:00Z Dual critical pre-authentication vulnerabilities in PHP core allow unauthenticated remote code execution across PHP-FPM, Apache mod_php, and embedded web appliances through heap cleanup corruption and cross-request worker state reuse. https://blog.badprompt.dev/posts/cross-tenant-rag-poisoning-multi-modal-jailbreaks/ https://blog.badprompt.dev/posts/cross-tenant-rag-poisoning-multi-modal-jailbreaks/ 2026-10-27T00:00:00Z Adversaries exploiting multi-modal document parsers in shared retrieval pipelines can inject visually imperceptible prompt vectors that hijack responses rendered to other enterprise tenants. https://blog.badprompt.dev/posts/weaponizing-mcp-tool-schemas-indirect-prompt-injection-rce/ https://blog.badprompt.dev/posts/weaponizing-mcp-tool-schemas-indirect-prompt-injection-rce/ 2026-10-26T00:00:00Z Adversaries can embed prompt injection control sequences directly into Model Context Protocol (MCP) tool description metadata, causing autonomous coding agents to execute arbitrary shell commands without user consent.