Most recent
navigate open esc close Corpus index built 2026-06-07 23:58 UTC

§ THE STACK / DATA LAYER

Model Hubs & Registries

HF Hub mirrors, ModelScope, BentoML

Vector stores, registries, memory, datasets: what the model knows and remembers.

What it is

Models live somewhere. The Hugging Face Hub is the public default; ModelScope (Alibaba) is the Chinese equivalent; Replicate hosts serverless fine-tunes; BentoML’s BentoCloud is the Python-native deployment registry. Most large operators run a private mirror of one of these so internal teams can pull models without going to the public internet, and so the operator’s own fine-tunes can be versioned and served through the same interface every framework already speaks.

What goes wrong

A private hub mirror is a webhook target, a model-weights store, and an API token issuer all in one. When the mirror is exposed without auth, an attacker pulls every model the operator has uploaded, including private fine-tunes that contain the operator’s training-data leakage. Worse, most mirrors implement the same /api/models/upload endpoint as the upstream, so an attacker can push a malicious model into the operator’s namespace and wait for an internal team to pull it. The supply-chain risk is real: PyTorch models execute arbitrary Python on load.

How we test

We probe /api/models (HF-compatible) or /api/v1/models (ModelScope) for the inventory and read the model card metadata. Model names plus sizes characterise the operator and identify private fine-tunes. We do not download weights. Where the upload endpoint is reachable we confirm reachability with an OPTIONS request and stop.

Receipts

Research

Every survey, case study, and disclosure we've published that touches this layer of the stack. Counts on the cells above tally these directly.

Queued

We haven't surveyed this category yet. The technology is on our map; the receipts will follow when the cross-cloud survey lands. Browse the research feed for what's already published, or watch this page.