api-web-search-with-citations

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

• Third-party content exposure detected (high risk: 0.85). The skill instructs the agent to read a public web page at runtime (“Read the source post first: https://claude.com/blog/web-search-api”); the fetched page body is outsider-authored free text that can be ingested into the LLM context via the web-search/URL-fetch workflow.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent to "Read the source post first: https://claude.com/blog/web-search-api", so the agent would fetch that external page at runtime and use its content to control prompts/instructions, making it a required runtime dependency.