ima-dai-sdk

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

• Third-party content exposure detected (medium risk: 0.30). In the HTML5/Web StreamManager workflow, the runtime reads timed metadata and stream event fields derived from the requested stream/manifest (e.g., HLS ID3 frames and DASH custom event payloads) and passes their raw data into streamManager.processMetadata(...), so outsider-authored content injected into those manifests/streams can become LLM-readable free text at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill loads remote JavaScript libraries at runtime (which execute code) via the script URLs https://imasdk.googleapis.com/js/sdkloader/ima3_dai.js and the protocol‑relative //www.gstatic.com/cast/sdk/libs/caf_receiver/v3/cast_receiver_framework.js and //imasdk.googleapis.com/js/sdkloader/cast_dai.js, so these external resources directly execute remote code the skill depends on.