hunt-agent-prompt-injection

Warn

Audited by Socket on Nov 2, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

The skill is a coherent but high-risk agent red-teaming guide. It contains no hidden installer, credential collection, or fixed exfiltration destination; its main risk is enabling actionable prompt-injection, data-exfiltration, and unauthorized-access testing without an explicit authorization boundary.

Confidence: 94%Severity: 78%
Audit Metadata
Analyzed At
Nov 2, 2026, 10:15 PM
Package URL
pkg:socket/skills-sh/elementalsouls%2Fclaude-agenthunter%2Fhunt-agent-prompt-injection%2F@45ab16cc268564882b2aed92745a51213fa6bcd2