google-shopping-light-search
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes product data from the Frevana API which is considered untrusted external content. * Ingestion points: The script
scripts/search_google_shopping_light.shfetches data fromai-factory.frevana.comand outputs it to stdout and local files. * Boundary markers: No explicit delimiters are provided in the skill instructions to separate search result content from agent instructions. * Capability inventory: The skill can execute local bash scripts, perform network requests viacurl, and write to the local file system. * Sanitization: The script performs structural JSON validation using Python, but does not sanitize the text content of the results. - [COMMAND_EXECUTION]: The skill executes a local bash script (
scripts/search_google_shopping_light.sh) to perform API requests and data processing. This is the intended behavior for the skill's search functionality. - [DYNAMIC_EXECUTION]: The bash script uses inline Python code blocks to handle JSON generation and validation. * Evidence: The script
scripts/search_google_shopping_light.shpipes Python code intopython3 -to process environment variables into a JSON payload and to validate the JSON structure of the API response.
Audit Metadata