copier

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly directs the agent/user to fetch templates from arbitrary Git repositories (see "Template sources" in SKILL.md: "copier copy gh:owner/repo", "git+https://...", etc.), and the reference docs show those templates can contain _tasks, _migrations, and _jinja_extensions that are executed when trusted, meaning untrusted, user-provided repo content is ingested and can materially influence tool actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill documents fetching copier templates at runtime via git URLs (e.g., git+https://github.com/owner/repo.git, gh:owner/repo, or git@github.com:owner/repo.git), and those remote templates can contain _tasks, _migrations, or _jinja_extensions that run commands or control prompts when trusted, so fetching those URLs at runtime is a clear high-confidence execution/control risk.