learn
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior, prompt injections, or unauthorized data access patterns were identified.
- [DATA_EXFILTRATION]: The skill accesses files within
~/.claude/learning/to retrieve user preferences and lesson history. This is a scoped use of the filesystem for application state and does not expose sensitive system information or credentials. - [COMMAND_EXECUTION]: The skill does not execute arbitrary system commands. In 'hands-on' mode, it asks the user to specify a directory and perform the tasks, ensuring the user remains in control of the environment.
Audit Metadata