architectural-specification
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: No instruction override, safety bypass, or system prompt extraction patterns were detected. The content remains consistently focused on the stated architectural purpose.
- [DATA_EXFILTRATION]: No network operations (curl, wget, fetch) or access to sensitive local file paths (like credentials or SSH keys) were found.
- [OBFUSCATION]: No encoded content, Base64 strings, zero-width characters, or homoglyph attacks were identified in the document.
- [REMOTE_CODE_EXECUTION]: The skill does not download external scripts, install packages, or attempt to execute any remote code.
- [COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or system-level operations present in the instructions.
- [DYNAMIC_CONTEXT_INJECTION]: No use of the load-time execution syntax (!
command) was found in the SKILL.md file.
Audit Metadata