aar
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of Markdown templates and instructions. No executable code (Python, Node.js, Shell) is present in the analyzed files.
- [REMOTE_CODE_EXECUTION] (SAFE): No remote code patterns, external downloads, or dynamic execution mechanisms (eval/exec) were found.
- [DATA_EXFILTRATION] (SAFE): No network operations (curl, wget, fetch) or sensitive file access patterns were detected. The skill does not communicate with external domains.
- [PROMPT_INJECTION] (SAFE): The instructions are focused on structured debriefing and do not contain attempts to bypass AI safety filters or override system instructions.
- [INDIRECT_PROMPT_INJECTION] (SAFE): While the skill ingests user input via the
$ARGUMENTSplaceholder to populate templates, it possesses no dangerous capabilities (like file writing or network access) that could be exploited via malicious data ingestion.
Audit Metadata