tool-install

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

• Third-party content exposure detected (medium risk: 0.30). The skill ingests outsider-authored text by fetching a user-provided URL/repo (including GitHub README/install scripts) via fetch_content and web_search, which exposes the agent to arbitrary injected content included in those pages.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch user-provided URLs at runtime and uses repository content (e.g., https://github.com/jesseduffield/lazygit) such as README/install scripts to shape the agent's research and generated prompts/plans, so external content can directly control the agent's instructions.