kpi-definition-helper
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTIONNO_CODE
Full Analysis
- [Indirect Prompt Injection] (HIGH): The skill processes untrusted external content (user-provided KPI definitions and requests) while possessing high-privilege capabilities including
Bash,Write, andEdit. - Ingestion points: User requests and KPI patterns provided in conversational context.
- Boundary markers: Absent. No instructions are provided to delimit data from instructions.
- Capability inventory:
Read,Write,Edit,Bash,Grep. - Sanitization: Absent. No filtering or validation of user-provided content is specified before passing to tools like
Bash. - [Command Execution] (MEDIUM): The skill explicitly allows the
Bashtool. While no malicious commands are hardcoded, the combination ofBashaccess with untrusted input processing (Category 8) poses a risk of command injection via the agent. - [No Code] (INFO): The skill contains only metadata and descriptive markdown with no functional scripts or implementation logic provided, making its actual behavior dependent on the base agent's interpretation of its capabilities.
Recommendations
- AI detected serious security threats
Audit Metadata