feedback
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to identify the host operating system using
echo "$OSTYPE"and to parse version information fromSKILL.mdfiles located within the~/.claude/skills/directory. - [COMMAND_EXECUTION]: The skill uses the
ghCLI or system commands (such asopen,xdg-open, orstart) to interact with GitHub. User-provided information is rewritten by the agent into a structured format before being passed as arguments to these commands, providing a layer of sanitization. - [DATA_EXFILTRATION]: Local environment metadata, including the OS type and versions of related skills, is collected and included in the feedback report. This data is transmitted to the developer's official GitHub repository (a well-known service) only after explicit user confirmation of the drafted issue.
Audit Metadata