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 Phase 1.3 workflow searches and processes the user's Gmail inbox via `mcp__google-workspace__search_gmail_messages` with a goal of inbox zero, so any outside sender can deliver free text into it without the agent selecting an item; WhatsApp capture review (Phase 1.4) and calendar event reads (Phase 3, which may include invite descriptions from external senders) are similar unsolicited feeds.
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.