web-search

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly performs web searches via tavily_search (scripts/search.py) and fetches arbitrary webpages with fetch(url) (scripts/fetch.py), and the SKILL.md workflow requires searching, fetching, and extracting content from public URLs—meaning untrusted third‑party content is ingested and can influence the agent's subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The scripts/fetch.py fetches arbitrary external URLs at runtime (e.g., python scripts/fetch.py "https://example.com") and returns the page content as Markdown intended for LLM consumption, allowing remote content to be injected into and directly control agent prompts.