Domain Evaluation Harness

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a configuration and implementation guide for an evaluation framework. It contains no executable code or malicious instructions.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a system for ingesting external data (task lists) to be processed by an agent, which is a known attack surface for indirect prompt injection.
  • Ingestion points: Task list JSON files (defined in Step 1).
  • Boundary markers: None specified in the provided shell script templates.
  • Capability inventory: Execution of agent logic via shell scripts (harness.sh) as defined in Step 2.
  • Sanitization: No sanitization logic is included in the templates, which is expected for a template-based guide.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 10:10 PM