api-monitor

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill monitors untrusted external sources (blogs, GitHub, social media) which could contain malicious instructions designed to manipulate the agent's output.
  • Ingestion points: External data sources defined in the create_scout query (changelogs, blogs, documentation).
  • Boundary markers: Absent; the query template does not use delimiters to wrap the external content.
  • Capability inventory: Limited to reporting and generating recommendations; no direct system-level write or execution capabilities are present in the skill definition.
  • Sanitization: No explicit sanitization or filtering of the ingested external content is defined.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 04:05 AM