AI Security Threat Feed

NotebookLM MCP: path traversal enables arbitrary file write

Joplin: stored XSS via Fountain renderer exposes notes

Joplin: path traversal in sync allows arbitrary file write

Joplin Desktop: update signature check skipped on Windows

CuteNews: authenticated file upload leads to RCE

Mistral Vibe: RCE via git hooks run before trust check

SiYuan: stored XSS in heading style enables Electron RCE

SiYuan: stored XSS in notebook names enables RCE

ToolHive: default network profile enables MCP lateral movement

lmdeploy: URL parser mismatch enables SSRF bypass

LMDeploy: eval() injection via malicious model config

vLLM: negative token ID crashes embeddings engine

SiYuan: SQL injection in getGraph exposes private notes

vLLM: memory leak in PD disaggregation causes DoS

Vvveb CMS: SSRF in oEmbed proxy leaks cloud metadata

Jupyter Server: Referer token leak in 5xx logs

Grafana Geomap: stored XSS enables Org Admin escalation

Grav CMS: callable-array bypass enables authenticated RCE

Grav CMS: callable guard bypass enables file read/write

amqp091-go: cleartext AMQP password retained in memory