troubleshoot-separation

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external diagnostic data.
  • Ingestion points: System logs, Method conditions, and Chromatograms (SKILL.md).
  • Boundary markers: Absent.
  • Capability inventory: Read, Grep, Glob, WebFetch, and WebSearch (SKILL.md).
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 07:13 AM