tdn-lookup

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

• Third-party content exposure detected (high risk: 0.93). SKILL.md:3 describes Tier 2/3/4/5 using WebFetch/Playwright to fetch and parse Confluence/TDN page bodies (outsider-authored public web content) and feed the extracted titles/excerpts/HTML-rendered text into the agent’s context as the “Usar diretamente” result.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly performs runtime fetches of external Confluence content (which is injected into the agent's context) via the REST API and site URLs such as https://tdn.totvs.com/rest/api/search?cql=<CQL_ENCODADO>&expand=body.view&limit=3 and https://tdn.totvs.com, so remote pages/JSON directly influence the agent's responses.