quality-assurance

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to drive a browser against arbitrary URLs (see "Browser trace — One-command capture" and the "Live web app QA" Explore phase which visits pages, extracts DOM links, and traces each link), meaning it fetches and ingests open/public (potentially user-generated) third-party content and uses those results to decide follow-up browsing/actions—exposing the agent to indirect prompt-injection risk.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs running the third‑party npm tool via "npx ecc-agentshield" (and even npx ecc-agentshield scan --opus --stream), which fetches and executes remote code at runtime and can run adversarial/agent workflows that control prompts—meeting the criteria for a risky runtime external dependency.