The Agent Skills Directory

[SAFE]: The skill does not contain any direct prompt injection attempts or instructions to bypass safety filters. All instructions are focused on providing empathetic and professional email drafting logic.
[DATA_EXPOSURE]: Data access is limited to internal context-gathering actions such as get_meetings, get_contact, and get_deal. These actions are necessary for the skill's functionality and do not involve hardcoded credentials, sensitive system file access, or unauthorized network operations.
[REMOTE_CODE_EXECUTION]: No remote code patterns, external script downloads, or dynamic code execution (like eval or exec) were found in the skill. It operates entirely within the platform's standard capability set.
[PROMPT_INJECTION]: While the skill ingests external data (meeting titles, contact names) via actions like get_meetings (Category 8 surface), the risk of indirect prompt injection is mitigated by the approval_required: true and auto_execute: false settings, ensuring a human reviews the generated draft before it is used. All capability usage is consistent with the primary purpose of the skill.

No-Show Follow-up