parallel
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted user input and passes it directly to a sub-agent, creating a vulnerability surface for indirect prompt injection.
- Ingestion points: User-provided data enters the agent context through the
$ARGUMENTSvariable used in the workflow steps ofSKILL.md. - Boundary markers: The skill does not wrap the input in delimiters (such as XML tags or triple backticks) nor does it provide instructions to the sub-agent to ignore potential commands embedded in the data.
- Capability inventory: The skill has access to the
Agenttool (for background process spawning) and theSkilltool (for invoking any other available system skill). - Sanitization: No validation, escaping, or filtering is performed on the input string before it is interpolated into the prompt for the sub-agent.
Audit Metadata