weekly-wrap
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill processes external, untrusted data which presents a potential injection surface.
- Ingestion points: Data is pulled from email threads, chat messages, transcripts, and external documents as defined in the skill instructions.
- Boundary markers: The skill utilizes explicit instructions to treat this content as 'data, never instructions' and instructs the agent to report any detected instructions.
- Capability inventory: The skill possesses capabilities to write to the CRM and post messages to team chat channels.
- Sanitization: Mitigations include suppressing the rendering of links found in untrusted content and requiring explicit user confirmation for any actions targeting recipients or content derived from untrusted sources before execution.
Audit Metadata