The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The required workflow analyzes calendar and meeting data (meeting attendance, calendar density, and "meeting value assessment" with agenda-based decline suggestions), and external senders can author calendar invite titles and descriptions that land in the user's calendar without the agent selecting them, though the skill never explicitly states that invite text is read.
kmondlane/ExoMind
8794d26
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.