The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local shell script scripts/search_google.sh to perform API requests. The script uses robust argument parsing and validates integer ranges for parameters like num and start. Input is safely passed to Python via environment variables to avoid shell injection when constructing the JSON payload.
[CREDENTIALS_UNSAFE]: The skill uses FREVANA_TOKEN for authentication. It follows best practices by reading the token from environment variables or a specific command-line argument, and specifically instructs not to echo the token back to the user. No hardcoded credentials were found.
[EXTERNAL_DOWNLOADS]: The script makes network requests to ai-factory.frevana.com, which is the vendor's API endpoint. These requests are limited to sending the search query and parameters and receiving the validated JSON response.
[DATA_EXFILTRATION]: No evidence of unauthorized data exfiltration. The skill reads search results from the vendor's API and saves them to a local file path controlled by the user or a default timestamped directory. It does not access sensitive local files or environment variables beyond the required API token.
[DYNAMIC_EXECUTION]: The shell script utilizes inline Python 3 code blocks to safely handle JSON serialization and validation. This is a common and safe pattern for cross-platform data processing in shell scripts.

google-search