penetration-testing

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The workflow directs the agent to scan targets, fetch exploit code, and scrape web/network services via command execution, exposing the LLM to outsider-authored text from target responses and downloaded exploits (low risk via active fetching).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill references external GitHub repositories via HTTPS (https://github.com/carlospolop/PEASS-ng/... and https://raw.githubusercontent.com/mzet-/linux-exploit-suggester/...) which are fetched and executed during privilege escalation checks, constituting weak third-party operational dependencies.