velt-comments-best-practices

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

• Third-party content exposure detected (high risk: 0.80). The skill's documentation and examples show the runtime consumption of user-generated comments/annotations from the Velt service (e.g., useCommentAnnotations(), VeltCommentsSidebar onCommentClick handlers and client.getCommentElement() flows in SKILL.md / AGENTS.full.md), and those annotations (untrusted third‑party content) are read and used to drive actions like navigation/seek/setLocation, which could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes a runtime script tag that loads and executes remote JavaScript from https://cdn.velt.dev/lib/sdk@latest/velt.js (used to initialize Velt in the HTML example), which is a required dependency fetched at runtime and therefore executes remote code—posing a runtime-executed external dependency risk.