illustration-ideas

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to browse and read arbitrary web pages provided via a URL (e.g., "You browse a page..." and the listed agent-browser commands like agent-browser open <url> / agent-browser get text @eX), which clearly ingests untrusted third‑party content that can influence subsequent illustration/decision actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly runs browser commands to fetch arbitrary pages (e.g., "agent-browser open " / "agent-browser get text @eX"), so the user-supplied URL ("") is retrieved at runtime and its content is injected into the agent context, thereby directly controlling prompts/instructions.