mapping-attack-techniques
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses inline Python scripts (
python3 -c) and CLI tools likerg(ripgrep) to query and transform local security data stored insecskills-core/ttp-index.json. These operations are transparent, handle JSON data safely, and are limited to the skill's local directory. - [EXTERNAL_DOWNLOADS]: The skill provides instructions for fetching external security advisories and vendor reports using the
defuddle.mdservice to convert content to Markdown for easier processing. - [DATA_EXFILTRATION]: Instructions for using the
defuddle.mdproxy include a clear security warning that URLs routed through the service are visible to a third party. The skill explicitly cautions against using this for sensitive adversary infrastructure or internal engagement URLs, demonstrating a defensive posture. - [PROMPT_INJECTION]: The skill's primary function involves ingesting and processing untrusted external data such as threat intelligence reports. This represents an indirect prompt injection surface. The skill manages this by instructing the agent to focus on extracting TTPs (Tactics, Techniques, and Procedures) rather than executing content, and provides a framework for human verification against established sources like the MITRE ATT&CK website.
Audit Metadata