The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is primarily composed of shell commands for performing network reconnaissance, enumeration, and exploitation of DNS services.
Evidence: Multiple code blocks in SKILL.md and references/dns-techniques.md provide bash commands for tools such as nmap, dig, nslookup, dnsrecon, dnsenum, fierce, and gobuster.
[DATA_EXFILTRATION]: The methodology explicitly includes Phase 6, which focuses on establishing DNS tunnels to bypass network security controls for data leakage or stealthy communication.
Evidence: Instructions and commands for using iodine, dnscat2, and dns2tcp are provided in both SKILL.md and references/dns-techniques.md to establish IP-over-DNS or command-and-control channels.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from remote DNS servers, creating an attack surface for indirect prompt injection via manipulated DNS records.
Ingestion points: Command output from dig, nslookup, and nmap probes in SKILL.md (Phases 1-5) and references/dns-techniques.md.
Boundary markers: None identified; the agent is instructed to analyze raw output directly.
Capability inventory: The skill utilizes extensive shell execution capabilities across all scripts to interact with network services.
Sanitization: No sanitization or validation of the remote DNS responses is implemented before the agent interprets the data.
[EXTERNAL_DOWNLOADS]: The skill references external tools and services for passive information gathering.
Evidence: Mentions querying well-known services like crt.sh (Certificate Transparency logs) and bgp.he.net (BGP routing information) for reconnaissance purposes.

dns-pentesting