signals-scout-health-checks
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it ingests untrusted data from external sources and possesses reporting and query capabilities.
- Ingestion points: The skill reads untrusted data from
health-issues-get(payload, title, and summary fields) andexecute-sql(retrieving details from thesystem.ingestion_warningstable). - Boundary markers: The instructions explicitly advise the agent to treat these fields as data to report rather than instructions, suggesting the use of quotes for untrusted snippets.
- Capability inventory: The skill can execute SQL queries (
execute-sql) and write to report channels (scout-emit-report,scout-edit-report). - Sanitization: The prompt contains a dedicated 'Untrusted data' section that warns the agent that these values can be set by external users and must never be treated as authorized commands.
Audit Metadata