The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting untrusted data from the user's Google Calendar.
Ingestion points: The scripts/gcal.py script retrieves and displays the summary, description, and location fields from calendar events across several functions (cmd_events, cmd_week, and cmd_json).
Boundary markers: The script outputs event data in a plain text bulleted list or JSON format without using specific delimiters or protective instructions that would warn the AI to treat the event content as data rather than instructions.
Capability inventory: This skill operates within a shell environment where the agent can execute commands and access local files, creating an exploitable surface if an event contains 'jailbreak' style instructions.
Sanitization: There is no filtering or sanitization of the retrieved event content before it is passed into the agent's context.

google-calendar