paper-fulltext-harvest

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

• Third-party content exposure detected (high risk: 0.90). The skill autonomously fetches and scrapes public third‑party web content (publisher pages and OA aggregators) — see references/browser-fallback.md (driving the user's Chrome to arbitrary publisher landing pages) plus clients.py and supplements.py (Unpaywall/OpenAlex/Crossref/DOI landing page requests and HTML parsing) — and the agent is expected to read/interpret that content to decide retries, extract PDF URLs, and drive follow‑up actions, which meets the criteria for exposure to untrusted third‑party content.