signals-scout-insight-alerts
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data sourced from the PostHog environment, creating a potential surface for indirect prompt injection.
- Ingestion points: The skill ingests data from
alerts-list,alert-get, andinsight-get, which includes user-defined alert names, conditions, and insight details (SKILL.md). - Boundary markers: There are no instructions for the agent to use delimiters or ignore instructions potentially embedded within the alert metadata or insight content.
- Capability inventory: The agent possesses write capabilities through the
emit_reportandedit_reporttools, which allow it to influence the project's inbox and report channel (SKILL.md). - Sanitization: The instructions do not specify any sanitization, validation, or escaping of the external content retrieved from the analytics tools before it is included in reports.
Audit Metadata