ln-71-operations-investigator

Pass

Audited by Gen Agent Trust Hub on Oct 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze operational evidence such as logs, traces, and metrics from external systems, which represents a potential surface for indirect prompt injection.
  • Ingestion points: External operational evidence including service logs, metrics, traces, and deployment history described in the 'Collect and Correlate Evidence' section.
  • Boundary markers: The instructions mandate redacting sensitive values and require the agent to 'distinguish facts, inferences, and unverified claims'.
  • Capability inventory: The workflow is explicitly restricted to 'read-only incident diagnosis' and forbids changing services, configuration, or credentials.
  • Sanitization: Includes explicit instructions to 'redact sensitive values' during telemetry queries.
  • [SAFE]: The skill contains strong safety boundaries, explicitly prohibiting any changes to live systems, configuration, or credentials. It uses a structured 'Execution contract' that requires evidence (PROVEN, CLEARED) rather than self-reported success, which helps maintain the integrity of the diagnostic process.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 18, 2026, 05:49 PM