Git

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (SKILL.md and commands.md) explicitly directs interacting with remote repositories (e.g., "git clone https://github.com/user/repo.git", "git submodule add URL", "git remote add origin URL"), so the agent will fetch and read untrusted public repository content as part of Git tasks, which could contain user-generated instructions that influence subsequent actions.