The Agent Skills Directory

[SAFE]: The skill implements robust safety mechanisms, including a mandatory human-approval gate for all calendar operations and a detailed execution plan displayed before any changes are made.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from calendar event titles and descriptions which could theoretically contain malicious instructions. However, the risk is mitigated by the requirement for human verification and the restricted scope of allowed actions (e.g., only modifying user-owned events).
[SAFE]: All operations are scoped to a specific user-defined week and authenticated calendar access, with explicit instructions not to touch 'untouchable' meetings or perform actions outside the approved plan.

calendar-defrag