protheus-business

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (SKILL.md Lookup Strategy) explicitly instructs the agent to perform WebSearch/WebFetch on the public TOTVS Developer Network (site:tdn.totvs.com) and to use a Playwright fallback to navigate and extract page text/screenshots, meaning it fetches and ingests untrusted public web content (TDN) that the agent will read and use to drive responses and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime lookup strategy explicitly performs WebSearch/WebFetch and a Playwright browser_navigate/browser_snapshot against the TOTVS Developer Network (https://tdn.totvs.com), fetching documentation at runtime that is extracted and injected into the agent's response context—i.e., external content directly influences prompts/results.