benchmark-methodology

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists exclusively of markdown text providing instructions and scoring rubrics. No code, scripts, or automation logic are included in the skill content.
  • [INDIRECT_PROMPT_INJECTION]: The benchmarking methodology involves the agent visiting and analyzing untrusted external content, such as competitor websites, LinkedIn profiles, and review directories.
  • Ingestion points: Competitor sites, LinkedIn, and review platforms (e.g., Clutch.co) as described in the 'How to collect the data' section.
  • Boundary markers: Absent. The skill does not explicitly instruct the agent to ignore directives potentially embedded in competitor website content.
  • Capability inventory: None. The skill itself does not request or use any high-privilege tools or execution capabilities.
  • Sanitization: Absent.
  • Note: While an ingestion surface exists, the risk is minimal as the skill lacks dangerous tools (shell access, network exfiltration) for an attacker to exploit via indirect injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 10:01 PM
Security Audit — agent-trust-hub — benchmark-methodology