dns-network

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

• Third-party content exposure detected (high risk: 0.80). Flagged because SKILL.md explicitly instructs running network queries against arbitrary public hosts and resolvers (e.g., dig @8.8.8.8 example.com, curl https://example.com, traceroute/mtr), meaning the agent would ingest untrusted third‑party DNS/web responses and use those outputs to decide next actions.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly includes commands that use sudo and modify system files (e.g., appending/removing /etc/hosts, running sudo tcpdump, and checking/modifying iptables), which direct the agent to change the host's state and require elevated privileges.