signals-scout-apm

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 external data in the form of OpenTelemetry spans, traces, and service attributes fetched via the apm-* toolset. It also has the capability to write to a report channel and a scratchpad.
  • Ingestion points: Untrusted telemetry data is ingested from the environment via tools like apm-trace-get, query-apm-spans, apm-spans-aggregate, and apm-attribute-breakdown (SKILL.md).
  • Boundary markers: The skill lacks explicit instructions or boundary delimiters to ensure the agent ignores potentially malicious content embedded within processed telemetry data, such as exception messages or span attributes.
  • Capability inventory: The skill utilizes report authoring tools (emit_report, edit_report) and state management tools (scout-scratchpad-remember) to persist its findings.
  • Sanitization: There is no evidence of content sanitization or validation of the fetched telemetry data before it is processed or included in reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 09:33 PM
Security Audit — agent-trust-hub — signals-scout-apm