The Agent Skills Directory

[SAFE]: The skill implements a structured workflow for prompt optimization based on recognized industry best practices. It correctly separates platform instructions from user-provided persona layers, minimizing the risk of prompt injection and ensuring consistent agent behavior.
[SAFE]: References to external sources (OpenAI, Anthropic, Google Gemini, and academic research) are used strictly for instructional context and documentation provenance. No dynamic remote code execution or untrusted downloads were found.
[SAFE]: Tool usage is addressed through policy-level instructions (when and how to use tools) rather than hardcoded schemas or executable scripts, maintaining a secure separation between the agent's logic and the underlying tool implementations.
[SAFE]: The skill promotes the use of environment-agnostic base prompts with minimal adapter layers for specific models, reducing the likelihood of cross-model behavioral failures or unintended security bypasses.
[SAFE]: No obfuscation, data exfiltration, privilege escalation, or persistence mechanisms were detected in the instructions or the accompanying reference files.

prompt-optimizer