meeting-prep-live
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Google Drive, Gmail, and Slack threads to generate meeting briefs. This creates a potential surface where malicious instructions embedded in those documents or messages could attempt to influence the agent's behavior during the synthesis process.
- Ingestion points: Data enters the context via the Google Calendar, Google Drive, Gmail, and Slack connectors as specified in the 'Execution' section.
- Boundary markers: The skill requires the agent to 'cite the real source for every claim' and 'mark anything you couldn't find as a gap', which provides a clear distinction between sourced data and agent synthesis.
- Capability inventory: The skill is limited to reading data through established connectors and generating an artifact; it does not have capabilities for arbitrary file writing, shell command execution, or non-platform network requests.
- Sanitization: No explicit sanitization or filtering of external message/document content is described in the instructions.
- [SAFE]: The skill explicitly instructs the agent to maintain user autonomy, listing 'sending anything without being asked' as an anti-pattern and advising the agent to 'offer the next action... but don't take it unasked'. This reduces the risk of the agent being manipulated into performing unauthorized tasks.
Audit Metadata