The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes integration workflows that ingest external data, such as SMS messages and CRM notes, which serves as a potential attack surface for indirect prompt injection.
Ingestion points: SMS engagement workflows using zoomphonesms:// and zp-sms-log-event, and user-provided call notes handled by zp-notes-save-event.
Boundary markers: The documentation encourages security hardening through origin validation and signature validation, but does not provide specific delimiters for LLM interaction within the provided code snippets.
Capability inventory: The skill utilizes standard REST APIs via fetch and cross-origin communication via postMessage.
Sanitization: The provided guidance includes verifying event.origin and validating webhook secret tokens to ensure data integrity.
[EXTERNAL_DOWNLOADS]: The skill references the official Zoom CRM sample repository on GitHub to provide architectural guidance for developers.
Evidence: Reference to https://github.com/zoom/CRM-Sample in references/crm-sample-validation.md.

phone