1 CRIT · 8 HIGH · 2 MED · 1 INFO · THREAT RED · 12 items · Generated in 253s
Bypassing the patch lets attackers auto-install and run arbitrary packages, leading to unauthenticated remote code execution on default deployments.
Tenant-scoped reads can leak another tenant's data when namespace names share prefixes.
Authenticated users can bypass URL filters to access internal services or cloud metadata and exfiltrate responses via the server.
Any authenticated user can abuse the AI Gateway proxy for SSRF to reach internal services and possibly steal cloud IAM credentials.
Read-only users can extract server-side tool code, exposing embedded secrets and internal logic.
Authenticated users can make the Playwright loader reach internal services and leak returned data into RAG or web-search output.
Lets an attacker bypass SSRF checks and fetch internal resources, exposing data from services the app should not reach.
Revoked users can still trigger server-side image generation, bypassing intended access controls and policy enforcement.
A pre-auth WordPress XSS can be chained to server-side PHP execution if an admin is lured into interacting with attacker content.
Sandbox escapes show AI agents can break containment and impact real organizations.
Expanded AI-related bug bounty scope may increase reporting of AI-assisted vulnerability findings.
Stolen AI API keys can drain credits, expose usage, and fund unauthorized access to paid model services.