The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data from web research and external enrichment tools in Step 2 of its execution flow. This establishes an indirect prompt injection surface where malicious content on a researched webpage could attempt to influence the drafted outreach. However, the skill defines strict output formats, style guidelines, and templates that mitigate this risk by providing a structured framework for the agent's response.
[SAFE]: The skill does not contain any evidence of credential harvesting, unauthorized network connections, or persistence mechanisms. Its functionality is consistent with its stated purpose of drafting sales outreach.

draft-outreach