working-with-scouts
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages AI agents (scouts) that ingest data from PostHog projects and external URLs. Ingestion points: Scouts process project events and logs, and can be configured for external network access as described in the delegation recipes. Boundary markers: The instructions explicitly advise users to tell scouts to treat fetched external data as untrusted content rather than instructions. Capability inventory: Scouts have the ability to read project data and, if granted specific write scopes, can modify PostHog objects such as notebooks, dashboards, and insights. Sanitization: The system includes a safety or actionability judge that can suppress reports before they reach the inbox.
- [SAFE]: No malicious patterns detected. All external resources and tools are official PostHog vendor assets, and the skill promotes secure configuration practices.
Audit Metadata