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

mobile-testing

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's audit process, described in references/placement.md, ingests untrusted data from Jira issue comments/descriptions and GitHub Pull Request bodies/comments to perform test layer classification. This ingestion of external data combined with the skill's capabilities to execute repository tooling and filesystem scripts creates an attack surface for indirect prompt injection. The skill lack explicit boundary markers or sanitization for this external input.
  • [DYNAMIC_EXECUTION]: The file references/placement/inventory.md contains shell commands that execute inline Python scripts using heredocs (`python3
  • <<PY`). While the logic in these scripts is static and limited to benign filesystem inventory (searching for test files and counting test cases), the pattern of executing code from within markdown files is a dynamic execution vector.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 07:59 PM
Security Audit — agent-trust-hub — mobile-testing