The Agent Skills Directory

[SAFE]: The skill consists of high-level analytical instructions focused on marketing performance diagnosis and experimental design. It does not invoke any external tools, system commands, or network operations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external marketing data (e.g., landing page content, ad copy), which inherently presents a surface for indirect prompt injection. However, the skill lacks any capabilities—such as automated file writing, network exfiltration, or code execution—that could be exploited via this vector. Furthermore, the instructions explicitly require human approval for high-risk actions like spend or publishing, providing a necessary control layer.

improve-conversion