recon-nmap

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

• Potentially malicious external URL detected (high risk: 0.80). The CI workflow includes a runtime command that fetches and pipes a remote script to the shell—"curl -s https://raw.githubusercontent.com/aquasecurity/tfsec/master/scripts/install_linux.sh | bash"—which downloads and executes remote code during skill runtime (the workflow also references external GitHub Actions that run code at runtime).

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The prompt repeatedly instructs the agent to run privileged nmap operations (explicit "sudo" examples) and techniques to bypass detection (packet fragmentation, decoys, source-IP spoofing), which push the agent to obtain elevated/raw packet privileges and evade security mechanisms on the host.