The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection surface identified where the agent ingests untrusted data from external DNS servers.
Ingestion points: DNS record values (TXT, MX, SPF, DMARC) retrieved via dig, nslookup, or host commands as described in SKILL.md and cheatsheet.md.
Boundary markers: Absent. There are no instructions or delimiters provided to signal that data from DNS records should be treated as untrusted content.
Capability inventory: The skill uses standard system binaries (dig, nslookup, host) for network diagnostics. No high-privilege capabilities or dynamic execution sinks were found.
Sanitization: Absent. DNS record outputs are processed and interpreted without specific filtering or sanitization.
[COMMAND_EXECUTION]: Provides detailed command-line examples for network diagnostic tools (dig, nslookup, host). These examples are standard for the tool's intended purpose of DNS troubleshooting and record verification.
[EXTERNAL_DOWNLOADS]: References documentation, RFC standards, and public DNS resolvers from well-known and trusted sources.
Evidence: Links to isc.org, rfc-editor.org, and public resolvers like Google (8.8.8.8) and Cloudflare (1.1.1.1).
Context: These resources are used neutrally for educational and diagnostic reference purposes.

bind-utils