Ghostcommit exposes the image blind spot in AI code review
ASSET Research Group's proof-of-concept hides a secret-stealing instruction inside a PNG that text-based reviewers skip.
By Ryan Merket · Published
Primary source: BleepingComputer
Why it matters
Ghostcommit turns a common AI coding workflow into a supply-chain risk: the PR looks clean because the instruction lives in a file reviewers skip, while the leak happens later during normal agent work.
ASSET Research Group and Sudipta Chattopadhyay published a proof-of-concept this week showing how a malicious pull request can hide a prompt-injection instruction inside a PNG, pass AI code review, and later induce a coding agent to leak a repository's .env secrets into source code.…