AI Security Threat Feed

Stanza: unsafe pickle in lemma model load enables RCE

node-forge: PKCS#12 MAC bypass via zero iteration count

MLflow: path traversal exposes server files via prompts

LiteLLM: file read via vertex_ai_credentials alias

Rembg: unauthenticated path traversal leaks files

vLLM: pickle deserialization RCE via multimodal cache

vLLM: image decoder abuse enables RCE/DoS

virtualenv: GBK quote bypass in xonsh activation enables RCE

virtualenv: prompt injection enables shell command execution

filelock: SQLite symlink redirect enables file overwrite

virtualenv: bash prompt injection allows command execution

SiYuan: WebSocket auth bypass leaks live event stream

obsidian-tc: HITL approval bypass runs destructive tools

Flexible GraphRAG: Cypher injection via FalkorDB adapter

virtualenv: shell injection in launcher path runs code

virtualenv: PS1 escaping flaw allows shell command execution

vLLM: unvalidated mm_processor_kwargs enables remote DoS

vLLM: RCE via code_revision in /tokenize endpoint

vLLM: forged multimodal UUID crashes inference engine

MLflow: path traversal allows arbitrary file write