nlweb-setup

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

• Third-party content exposure detected (medium risk: 0.65). SKILL.md instructs the runtime setup to “Fetch live docs first” from GitHub (public web content) and to “Web-search … release_notes” for the most recent release note; if the skill later ingests those fetched README/docs/release-note texts into the agent’s LLM context, that is outsider-authored free text via runtime HTTP fetches.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill instructs fetching and cloning the NLWeb repository (https://github.com/nlweb-ai/NLWeb and its docs/files like config/prompts.xml) as part of setup (git clone + install + run), which will fetch and execute remote code and provide prompt/config files that directly control runtime agent behavior.