CacheDeception

Warn

Audited by Socket on Nov 16, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

The skill is coherently focused on web cache exploitation, with no evident credential exfiltration, malicious installer, or command execution. Its main concern is operational: it gives an AI agent detailed instructions for high-impact offensive testing, including retrieving victim data and poisoning shared caches, without explicit per-action approval or clear safeguards restricting tests to authorized, controlled accounts. Classify as suspicious/high risk for autonomous exploitation capability, not confirmed malware.

Confidence: 93%Severity: 76%
Audit Metadata
Analyzed At
Nov 16, 2026, 12:02 PM
Package URL
pkg:socket/skills-sh/security-talent%2Fbugskill-ai%2Fcachedeception%2F@74724abf8cf283175a67a223e338aad32b8bbb6cce026413680fb3945da98d2a