mercadolivre-search

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The script explicitly extracts browser cookies (including authentication cookies), caches and can export them—representing credential-access/exfiltration risk—even though there is no obfuscated code, hidden exec/rce, or obvious remote backdoor in the repository.

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

• Third-party content exposure detected (high risk: 0.90). This skill directly fetches and parses public Mercado Livre web pages (see SEARCH_URL/BASE_URL and MercadoLivreClient.fetch_html in scripts/mercadolivre.py) — including product pages, user-generated reviews and descriptions via the search/details/reviews/seller/categories commands and the --url argument — so it ingests untrusted third-party content that could carry indirect prompt injections.