product-enrich
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted product metadata from external CSV files and user input to generate classifications and tags. It lacks specific instruction boundary markers or sanitization logic, creating a vulnerability surface where malicious data could influence agent behavior.
- Ingestion points: The skill reads product data from
./products.csv,product-library.csv, and user-pasted text inputs. - Boundary markers: Absent; the instructions do not define delimiters or specific "ignore embedded instructions" warnings for the processed product rows.
- Capability inventory: The skill utilizes tools including
Read,Write,Edit,Bash,Glob,Grep, andAskUserQuestion(defined inSKILL.md). - Sanitization: Absent; there is no defined process for escaping or validating product field values before they are analyzed by the model.
Audit Metadata