The Agent Skills Directory

[SAFE]: No security issues were identified. The skill's file access is limited to its own configuration and brand data directories (~/.claude-marketing/) to provide contextually relevant PR advice. It does not attempt to execute remote code, exfiltrate credentials, or bypass security constraints.
[SAFE]: The skill processes user-provided descriptions of incidents to generate response plans. While this represents a surface for potential indirect instructions, the skill lacks dangerous capabilities such as network access, file writing, or command execution, rendering the risk negligible.

crisis-response