conversion-funnel-anomaly-detector
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests structured data from external files, which constitutes a surface for indirect prompt injection attacks.
- Ingestion points: The
run.pyscript reads data from an arbitrary file path provided via the--inputargument. - Boundary markers: The script lacks clear delimiters or instructions for the AI agent to ignore potentially malicious instructions embedded within the input JSON/CSV.
- Capability inventory: The skill possesses filesystem read and write capabilities through the
pathlibmodule inrun.py. - Sanitization: No input validation or sanitization is performed on the data payload before it is parsed and processed.
Audit Metadata