Database/AI/ML frameworks & serving
PyTorch (`weights_only` unpickler): Bypass of the `weights_only` allowlist
CVSS 8.8CVE-2026-24747AI/ML frameworks & servingcurated
Impact
Bypass of the weights_only allowlist → arbitrary code execution
Who can reach it
Customer-supplied checkpoint; defeats the mitigation shipped for CVE-2025-32434
What to do
Rebuild images on torch >= 2.10.0. Proves the format itself, not the flag, is the control — push checkpoint-format policy (safetensors-only ingest) rather than version-chasing
References
Related entries
- vLLM (hardcoded `trust_remote_code`, second instance): Same class, two more model files, through 0.18.0CVE-2026-27893 · vLLM (hardcoded `trust_remote_code`, second instance)High
- JupyterHub OAuthenticator: Authenticated user bypasses the intended identity checkCVE-2026-33175 · JupyterHub OAuthenticatorHigh
- LiteLLM (`/config/update`): Endpoint does not enforce admin authorizationCVE-2026-35029 · LiteLLM (`/config/update`)High
- BentoML (Dockerfile generation): Injection into generated DockerfileCVE-2026-35044 · BentoML (Dockerfile generation)High
- Jupyter Server: Path traversal in the REST APICVE-2026-35397 · Jupyter ServerHigh
- LiteLLM (`/guardrails/test_custom_code`): RCE via bytecode rewritingCVE-2026-40217 · LiteLLM (`/guardrails/test_custom_code`)High
This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.