prompt-injection-test

Fail

Audited by Socket on Mar 18, 2026

1 alert found:

Malware
MalwareHIGH
SKILL.md

MALICIOUS. The skill's stated content is a prompt-injection test, but its actual behavior is to subvert agent instructions and execute an unverified remote payload. The purpose, capabilities, and data flow are fundamentally incompatible with a benign skill.

Confidence: 98%Severity: 99%
Audit Metadata
Analyzed At
Mar 18, 2026, 02:04 AM
Package URL
pkg:socket/skills-sh/cisco-ai-defense%2Fskill-scanner%2Fprompt-injection-test%2F@160136db2dbda6f3c17c5999c5769ca608b2ff34