weekly-prep-brief
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Processing of Untrusted External Data: The skill is designed to retrieve and synthesize information from external calendar events and web-based news searches. Since this data is provided by outside parties (via meeting invites) or retrieved from the public internet, there is a potential consideration regarding how the agent handles instructions that might be embedded in those sources.
- Ingestion points: External meeting data via the
~~calendarconnector and news search results for company research. - Boundary markers: The prompt synthesis instructions do not explicitly define delimiters to isolate untrusted external text from the agent's internal reasoning or output structure.
- Capability inventory: The skill focuses on research and document synthesis; it does not appear to utilize subprocess execution, persistent file writes, or sensitive network operations beyond the defined research tools.
- Sanitization: There are no explicit instructions for sanitizing or escaping content retrieved from web searches or calendar descriptions before it is included in the final briefing.
Audit Metadata