The Agent Skills Directory

[DATA_EXFILTRATION]: The skill transmits user search queries and a sensitive authentication token (FREVANA_TOKEN) to the external endpoint https://ai-factory.frevana.com/service/walmart-search.
[COMMAND_EXECUTION]: The skill relies on the execution of scripts/search_walmart.sh, which invokes shell commands including curl for network operations and mkdir/cp for file system manipulation.
[DYNAMIC_EXECUTION]: The script uses python3 - to execute inline Python code for generating JSON request payloads and validating the integrity of API responses.
[INDIRECT_PROMPT_INJECTION]: The skill ingests search results from an external API, which may contain malicious instructions designed to influence the agent's subsequent behavior.
Ingestion points: The script scripts/search_walmart.sh reads and returns data from the Frevana API endpoint.
Boundary markers: No specific delimiters or instructions to ignore embedded content were found in the skill definitions.
Capability inventory: Includes network access via curl and the ability to write files to user-specified paths via the --output parameter.
Sanitization: The skill performs JSON structural validation but does not sanitize or filter the content of the product listings before presenting them to the agent.

walmart-search