The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The script scripts/search_google_trends.sh uses curl to make POST requests to https://ai-factory.frevana.com/service/google-trends to retrieve search trend data based on user keywords.
[COMMAND_EXECUTION]: The skill requires the agent to execute a local Bash script (scripts/search_google_trends.sh) which in turn invokes curl and python3 to process data and interact with the network.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external API and provides it to the agent for summarization, creating a vulnerability surface.
Ingestion points: The scripts/search_google_trends.sh script fetches JSON data from the ai-factory.frevana.com endpoint.
Boundary markers: Absent. Neither the skill instructions nor the script provide delimiters or instructions for the agent to ignore potentially malicious content embedded in the API response.
Capability inventory: The skill can execute shell scripts, make network requests via curl, and write data to arbitrary local file paths via the --output parameter.
Sanitization: The script validates that the API response is valid JSON, but it does not perform any sanitization or filtering of the text content within that JSON.

google-trends