summarize

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow and reference docs (SKILL.md and references/cli.md) explicitly instruct summarizing arbitrary public URLs, YouTube links, and feeds (e.g., summarize "https://example.com", summarize "https://youtu.be/<id>", summarize "https://feeds.npr.org/..."), meaning the agent fetches and interprets untrusted third‑party content as part of its operation, which could allow indirect prompt injection to influence decisions or tool use.

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

• Potentially malicious external URL detected (high risk: 0.90). The summarize CLI fetches arbitrary web pages at runtime (e.g., summarize "https://example.com"), injecting remote content into the model context so external URLs can directly control the agent's prompts/outputs and are required for the skill to operate.