picklescan & ModelScan denylist gap: io.open (== builtins.open) via the un-denylisted io module

Benign security-research PoC for the huntr Model File Format program. The payload only writes a nonce marker file (echo / io.open) — no real harm, no network, no secrets (huntr §benign).

Vulnerability

Both scanners denylist builtins.open but miss io.open (the identical object) reached via the un-denylisted io module -> arbitrary file write at load. CLEANEST bypass: both scanners exit 0 with no error. Siblings: bz2/gzip/lzma.open, sqlite3.connect.

Measured evidence (2026-07-24)

Against picklescan 1.0.5 + ModelScan 0.8.8: both fail to flag; on load the benign marker fires. Re-verify: python -m nomos_redteam scanbypass.

Reproduce

import torch  # or: import pickle
obj = torch.load("model.pt", weights_only=False)   # fires the benign marker at load

Scan first to confirm the bypass:

picklescan -p model.pt      # exit 0 / no dangerous global
modelscan -p model.pt        # no issue reported
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support