benchmark

ART (Attacker-Reachable Sink Triage)

An open-source benchmark suite that pairs vulnerable and patched versions of the same function to test whether a model's verdict changes once the security control is added. It is published under MIT with a frozen dataset and hosted Kaggle tasks.

Known aliases

art-label-triage
Attacker-Reachable Sink Triage

Relationships

No evidence-backed relationships are recorded.

Current clusters

build1 publisher

Haiku labels three of eight patched twins as reachable vulnerabilities

A Kaggle-challenge benchmark called ART scores models on whether they still flag a function after the fix is applied. On eight synthetic pairs, the difference between price tiers showed up only on the patched half.

Publishers:dev.to

Reality

Evidence47
Adoption12
Hype gap−5
Incentives58
Confidence44