The Agent Skills Directory

[SAFE]: The skill consists entirely of Markdown documentation and reference files aimed at teaching users and AI agents how to structure system prompts. It does not include any executable scripts, binary files, or active command-line operations.
[DATA_EXPOSURE]: The skill includes an example of a 'Data Privacy & Security Aware Agent' in references/examples.md which explicitly provides guidance on detecting PII and applying anonymization techniques. It does not attempt to access or exfiltrate sensitive files like SSH keys or AWS credentials.
[PROMPT_INJECTION]: While the skill contains instructional language such as 'IMPORTANT' or 'CRITICAL', these are used legitimately to highlight formatting requirements (e.g., brace escaping for Python templates) and do not attempt to bypass agent safety filters or override system instructions.
[COMMAND_EXECUTION]: The documentation mentions tools like python_repl and bash_tool in examples to illustrate how an agent might use them, but the skill itself does not invoke these tools or execute any shell commands.
[EXTERNAL_DOWNLOADS]: The files mention standard development tools such as pytest and ruff within example scripts (init.sh), but the skill does not perform any automated downloads or package installations.

system-prompt-writer