create-new-skill
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to perform filesystem operations, specifically creating symbolic links (ln -s) to install newly scaffolded skills into the agent's configuration directory. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8). It ingests untrusted user input via the
AskUserQuestiontool and incorporates this data into generatedSKILL.mdfiles using theWritetool. There are no boundary markers or sanitization steps implemented to prevent potentially malicious instructions from being persisted in the new skill files. - [COMMAND_EXECUTION]: The reference documentation included with the skill provides guidance on implementing dynamic context injection (
!command) and lifecycle hooks, both of which facilitate the execution of arbitrary shell commands at runtime.
Audit Metadata