signals-scout-logs
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes log data which is an untrusted external source. An attacker could potentially embed malicious instructions within logs to influence the agent's behavior.
- Ingestion points: Untrusted data is ingested via tools such as
query-logs,logs-patterns-diff, andlogs-attributes-list. - Boundary markers: The instructions explicitly command the agent to treat all log content (message bodies, service names, templates, and attribute values) as "quoted data" to be analyzed, never as instructions or authorization for tool calls.
- Capability inventory: The agent has capabilities to write to the report channel (
emit_report,edit_report) and modify a persistent scratchpad (scout-scratchpad-remember). - Sanitization: The skill mandates a protocol requiring "independent corroboration from a separate read" (e.g., service aggregation or error-tracking cross-checks) before performing any write operation, preventing the agent from acting solely on suspicious or injected log content.
Audit Metadata