agora

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

• Third-party content exposure detected (high risk: 0.90). This skill's workflow explicitly instructs the agent to perform Level-2 doc-fetching and live fetches from public third‑party sites (e.g., https://docs-md.agora.io/* OpenAPI/docs, https://raw.githubusercontent.com/* installers, and GitHub quickstart repos) and says "Always fetch these; do not answer from memory", so the agent will ingest and act on untrusted public web content as part of its required workflow.

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

• Potentially malicious external URL detected (high risk: 0.90). The ConvoAI quickstart enforces cloning and running the official sample repo (git clone https://github.com/AgoraIO-Conversational-AI/agent-samples.git), which fetches remote code that will be executed locally and is required as the baseline dependency.