walmart-product-reviews
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of Walmart product reviews retrieved from the Frevana API. \n
- Ingestion points: The API response is processed within
scripts/search_walmart_product_reviews.sh.\n - Boundary markers: No explicit delimiters are used to wrap the reviews in the output returned to the agent context.\n
- Capability inventory: The skill has file-write capabilities (saving results to a specified path) and network access (calling the Frevana API).\n
- Sanitization: The script uses Python's
json.loads()to ensure the API response is valid JSON before processing it further.\n- [DYNAMIC_EXECUTION]: The shell scriptscripts/search_walmart_product_reviews.shgenerates and executes Python code using heredocs to construct request payloads and validate responses. The embedded Python code is static and used for standard data processing tasks.\n- [COMMAND_EXECUTION]: The skill executes shell commands, includingcurlfor network requests andpython3for JSON manipulation. These operations are scoped to the skill's primary purpose of data retrieval.
Audit Metadata