'Ghostcommit' hides prompt injection in images to fool AI agents, steal secrets

0 members (), 3,669 guests, and 19 robots.

Key: Admin, Global Mod, Mod

Comment Guidelines: Do post respectful and insightful comments. Don't flame, hate, spam.

KSC PAD 39A

KSC PAD 39B

TLE DATA

IDB.COM.AU
For all your TLE downloads.

Today's Birthdays

There are no members with birthdays on this day.

AUS-CITY Latest Photos

Angels? Light-beings?
by Alisa, August 8

Popular Topics(Views)

AUS-CITY Earthquake Map

Volcano earthquake report for Tuesday, 10 November 2026

• M -0.6 - 63 km ENE of Pedro Bay, Alaska

• M 0.6 - 47 km ENE of Adak, Alaska

• M 2.3 - 54 km ENE of Susitna North, Alaska

• M 1.4 - 60 km WNW of Nikiski, Alaska

• M 1.1 - 109 km S of McCarthy, Alaska

AUS-CITY Advertisements

Print Thread

Rate Thread

'Ghostcommit' hides prompt injection in images to fool AI agents, steal secrets #824684 Sat 11 Jul 2026 09:03:AM
Joined: Feb 2001 Posts: 800,656 Los Angeles, USA Webmaster Launch Director
Webmaster Launch Director Joined: Feb 2001 Posts: 800,656 Los Angeles, USA	'Ghostcommit' hides prompt injection in images to fool AI agents, steal secrets A PNG hiding a prompt injection could steal your repo's secrets, researchers demonstrate. The technique, dubbed 'Ghostcommit,' slipped past AI code reviewers CodeRabbit and Bugbot, which never open image files at all, then convinced a coding agent to read a repo's .env and write every secret into the code as a list of numbers. [...] https://www.bleepingcomputer.com/news/security/ghostcommit-hides-prompt-injection-in-images-to-fool-ai-agents-steal-secrets/