The Agent Skills Directory

[SAFE]: The skill's primary function is to assist users in refining AI prompts through a structured workflow of intake, clarification, and template-based generation.
[EXTERNAL_DOWNLOADS]: All external URL references point to official OpenAI documentation and cookbooks (platform.openai.com, cookbook.openai.com, developers.openai.com), which are trusted sources for prompt engineering guidelines.
[PROMPT_INJECTION]: The skill was evaluated for prompt injection vulnerabilities. While it processes untrusted user data (the raw prompts to be improved), it does not execute these prompts within its own environment. Instead, it transforms them into structured XML formats. It includes instructions to preserve user intent and avoid making guesses when information is missing.
[REMOTE_CODE_EXECUTION]: No patterns of remote code execution or shell command execution were found. The skill operates purely through instruction-based text processing.
[DATA_EXFILTRATION]: No network calls to third-party domains or sensitive file access patterns (e.g., .ssh, .aws) were detected. The workflow is entirely contained within the interaction loop.
[COMMAND_EXECUTION]: The skill does not utilize shell commands or perform file system modifications.

prompt-improvement