eventkit
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and processing of data from Calendar events and Reminders, which are external sources capable of containing untrusted or adversarial instructions.
- Ingestion points: Untrusted data enters the agent's context via
eventStore.events(matching:)andeventStore.fetchReminders(matching:)calls detailed inreferences/core-implementation.mdandreferences/eventkit-patterns.md. - Boundary markers: The instructions do not define boundary markers or provide explicit prompts to the agent to ignore or isolate instructions that may be embedded within event titles, notes, or reminder content.
- Capability inventory: The agent is granted the capability to modify the user's data store through
eventStore.save()andeventStore.remove()methods, as shown in the implementation examples inreferences/core-implementation.mdandreferences/eventkit-patterns.md. - Sanitization: There is no documentation or guidance on sanitizing, escaping, or validating the content retrieved from the event store before it is used in downstream logic or prompt interpolation.
Audit Metadata