THREAT OPS › Threat News › GuardFall: a universal shell injection vulnerability in open-source AI agents
GuardFall: a universal shell injection vulnerability in open-source AI agents
<p><strong>The filter that’s supposed to stop dangerous commands in AI coding agents and computer agents is the reason teams feel safe enough to skip the human checks. Yet a thirty year old shell trick walks straight through this filter. One agent design shows the fix is known.</strong></p> <p>AI coding agents and computer use agents run shell commands with your full account authority: your
Attributed threat actors
- SilenceG0091
MITRE ATT&CK techniques
Indicators of compromise
- 81cd67829191beba42eb54589375670122e1b57bsha1
- https://danusminimus.github.io/posts/Prompt-Injection-Leads-To-RCE-And-Sandbox-Escape-In-Antigravity/url
- https://kb.cert.org/vuls/id/221883url