The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access detected. The skill's logic is entirely focused on design quality and accessibility standards.
[SAFE]: The skill implements protective "Risk Gates" in references/operating-checklist.md. These gates categorize operations by risk level and mandate human-in-the-loop approval for sensitive actions like modifying DNS settings, importing contacts, or initiating live sends.
[SAFE]: The instructions in SKILL.md emphasize separating analysis from live-system actions, providing a defensive layer against accidental misconfiguration or automated errors.
[PROMPT_INJECTION]: While the skill processes untrusted external content (HTML/CSS templates), it is categorized as low risk due to the explicit instructions to treat design artifacts as data for audit rather than executable instructions, and the presence of the aforementioned risk gates.

emaildesignskill