dns-attacks
Fail
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONOBFUSCATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides links and references to external security repositories and documentation, such as hacktricks.xyz and the dnscat2 GitHub repository. These are standard cybersecurity tools and resources.
- [COMMAND_EXECUTION]: The documentation describes the use of network tools like dig, arpspoof, and dnsrecon. These tools are commonly used for infrastructure auditing and occasionally require elevated permissions (sudo) to manage raw sockets, which is typical for offensive security tooling.
- [DATA_EXFILTRATION]: Educational examples of data exfiltration over DNS are provided, such as encoding data as subdomain labels. These scripts are provided for testing visibility and detection mechanisms and do not represent an automated exfiltration capability.
- [OBFUSCATION]: The guides contain examples of how attackers hide malicious domains using homoglyph substitution (e.g., аpple.com using a Cyrillic character) and zero-width spaces (e.g., apple.com). These instances are clearly defined as attack patterns for study rather than functional obfuscation of the skill's own content.
- [INDIRECT_PROMPT_INJECTION]: The skill describes tools that ingest external DNS data, which provides a conceptual surface for indirect prompt injection. 1. Ingestion points: Output from dig and dnsrecon tools. 2. Boundary markers: Absent in the example shell scripts provided for lab use. 3. Capability inventory: Access to the Bash shell and file system via the agent. 4. Sanitization: The instructional scripts do not perform explicit sanitization of external DNS responses, reflecting a focus on raw data analysis.
Recommendations
- Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata