The Agent Skills Directory

[PROMPT_INJECTION]: The SKILL.md file includes detailed instructions in the 'Grok Visual Content Policy' and 'Grok Text Content Policy' sections that describe how to bypass the content moderation and safety filters of the underlying AI models through specific phrasing and 'artistic framing' techniques.
[COMMAND_EXECUTION]: The examples/python_wrapper_example.py script utilizes the subprocess.run() method to trigger shell commands. While used for automation in the provided examples, this pattern represents a risk for command injection if integrated with unvalidated external input.
[CREDENTIALS_UNSAFE]: Instructions in the Quick Start section of SKILL.md encourage users to store API keys in environment files (~/.secrets) and verify them using the echo command, which can lead to sensitive credentials or partial keys being recorded in the shell's command history.

omni-image