spark-recipe-invitation-manager
Pass
Audited by Gen Agent Trust Hub on Apr 30, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill ingests untrusted content from email threads, creating an indirect prompt injection surface.
- Ingestion points: The skill retrieves external email data using the
spark threadcommand inSKILL.md. - Boundary markers: No explicit delimiters or instructions are used to distinguish between meeting details and potentially malicious embedded commands.
- Capability inventory: The skill uses
spark emails,spark thread,spark events, andspark availabilityfor read-only access to user communication and schedule. - Sanitization: No validation or sanitization is applied to the email content before the agent generates summaries and recommendations.
Audit Metadata