signals-scout-first-report
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from several external sources including error tracking, logs, traces, and warehouse data. This data could contain malicious instructions designed to influence the agent's reporting behavior.
- Ingestion points: The skill instructions in SKILL.md specify fetching data from 'Error tracking', 'Logs, traces, and metrics', 'Product analytics', 'AI observability', and 'Session recordings'.
- Boundary markers: There are no explicit instructions or delimiters defined to isolate the untrusted data or to command the agent to ignore any embedded instructions within these data streams.
- Capability inventory: The skill uses the
emit_reporttool to publish findings and possesses extensive read capabilities for database, warehouse, and log sources. - Sanitization: The instructions do not define any sanitization, escaping, or validation of the content retrieved from external logs or traces before it is processed by the agent or included in the final report.
Audit Metadata