The Agent Skills Directory

[SAFE]: The skill provides legitimate instructions for desktop app automation using developer-sanctioned tools and protocols. No malicious patterns, obfuscation, or unauthorized data access were found.
[PROMPT_INJECTION]: The skill interacts with data from external applications, which is an inherent surface for indirect prompt injection. Ingestion occurs via agent-browser snapshot and get text (SKILL.md). No specific boundary markers or sanitization steps are provided, though capabilities are limited to UI interactions within the automated applications.

electron