open-gstack-browser

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

• Third-party content exposure detected (high risk: 1.00). The skill launches an AI-controlled Chromium and explicitly runs browse commands (e.g., Step 4: "$B goto https://news.ycombinator.com") and a sidebar agent that "reads content" and executes actions (Step 5), which means it fetches and interprets arbitrary public web pages/user-generated content as part of its workflow, allowing indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill downloads and executes a remote installer at runtime via curl -fsSL "https://bun.sh/install" (saved to a temp file and run with bash), which fetches and runs remote code as a required setup dependency — this directly executes external code, so it's flagged.