maf-online-endpoint

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill tells the agent to ask users for workflow-specific variables (including API keys), then render deployment YAMLs that may contain API keys in plaintext, which requires the LLM to receive and embed secret values verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). The runtime scoring path assets/score.py::run(raw_data) ingests outsider-authored free text from the request body (raw_data → json.loads → data.get("text")) and passes it to workflow.run(text).

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

• Potentially malicious external URL detected (high risk: 0.70). The deployment references the remote container image mcr.microsoft.com/azureml/openmpi4.1.0-ubuntu22.04:latest (assets/deployment.yml and scripts/deploy_sdk.py), which will be pulled and executed by the managed endpoint at runtime making it a required external artifact that executes remote code.