tier-judge

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill analyzes untrusted data from git diffs, presenting a surface for indirect prompt injection.
  • Ingestion points: Diff content parsed from git history or provided as context (SKILL.md).
  • Boundary markers: The instructions do not define delimiters or markers to isolate the content being analyzed from the agent's logic.
  • Capability inventory: The skill is restricted to classification and reasoning; it does not invoke tools for filesystem modification, network requests, or shell command execution.
  • Sanitization: No evidence of sanitization or escaping of the input diff data is found.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 09:33 PM