skills/metamask/skills/ab-testing/Gen Agent Trust Hub

ab-testing

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the ripgrep (rg) utility to search the project codebase for specific A/B testing hooks and feature flag configurations.
  • [COMMAND_EXECUTION]: The skill executes local scripts located in the project's .agents/skills/ directory using Node.js and Bash to perform compliance checks on code changes.
  • [PROMPT_INJECTION]: The skill's workflow involves reading repository files, which constitutes a surface for indirect prompt injection. This risk is inherent to code-review tasks and is mitigated by the restricted use of local tools for specific compliance purposes. Ingestion points: repository source files in directories like app, shared, and ui. Boundary markers: none explicitly defined. Capability inventory: local file search (rg) and script execution (node, bash). Sanitization: relies on fixed internal script paths and standard developer environment tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 06:10 PM
Security Audit — agent-trust-hub — ab-testing