The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/map_injectived_cli.py uses the subprocess.run function to execute the injectived binary. This is used solely to programmatically traverse the CLI's help menus and generate a command reference map. The implementation uses a list-based command argument structure, which is a secure practice to prevent shell injection.
[EXTERNAL_DOWNLOADS]: The skill instructions reference installing the injective-core package via npm. This is the official distribution method for the author's command-line tools and is considered a safe and standard dependency for the skill's intended use case.
[CREDENTIALS_UNSAFE]: The skill provides documentation on managing wallet passphrases and Ethereum private keys. It proactively identifies sensitive operations as 'unsafe' (consistent with the CLI's own naming) and includes clear guidance for the agent to maintain security, such as using restricted file permissions (mode 600) and avoiding the logging of secrets.

injective-cli