ljg-structure
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill does not contain any malicious patterns such as prompt injection, data exfiltration to external servers, or obfuscated code.
- [COMMAND_EXECUTION]: The skill employs dynamic context injection via shell commands to generate precise timestamps using the
dateutility. These timestamps are used to create unique filenames and document metadata for organization. This is a legitimate use of the platform's capabilities for productivity. - [DATA_EXPOSURE]: The skill interacts with the local file system by writing analysis results to
~/Documents/notes/and subsequently reading them back to confirm the operation. This behavior is transparent and aligns with the skill's functional description as a structural analysis and note-taking tool. - [PROMPT_INJECTION]: The skill processes user-supplied information but includes specific workflow steps (e.g., paraphrasing, identifying motifs, and internal 'wind tunnel' simulations) that act as logical boundaries, reducing the likelihood of accidental instruction override.
Audit Metadata