The Agent Skills Directory

[SAFE]: The skill demonstrates secure practices for managing sensitive information by instructing developers to retrieve API tokens via Application.get_env or runtime configuration rather than hardcoding credentials.
[EXTERNAL_DOWNLOADS]: The skill references the req package from Hex (the official Elixir package registry). This is a well-known, standard library for the ecosystem and is used as intended for its primary purpose.
[COMMAND_EXECUTION]: Includes standard Elixir function calls for network operations and file streaming (File.stream!). These are restricted to the primary purpose of making HTTP requests and handling response data safely.
[DATA_EXFILTRATION]: While the skill involves network operations, it uses placeholder URLs (example.com) and encourages the use of configured base URLs from application environment variables, posing no risk of unauthorized data transfer.
[PROMPT_INJECTION]: No evidence of instructions attempting to bypass safety filters or override agent behavior was found.
[INDIRECT_PROMPT_INJECTION]: The skill establishes a data ingestion surface by processing external API responses. However, it mitigates risks by demonstrating robust pattern matching on HTTP status codes and centralized response handling, reducing the likelihood of unexpected data influencing application flow.

req-http-client