datadog-cli

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow (e.g., logs tail/search/patterns/trace/context, metrics query, and dashboards get/update) reads free-text fields contained in Datadog logs/dashboards that can be submitted by an outsider into the monitored Datadog sources (via their application logs/events or dashboard metadata), and the workflow streams/returns that content to the agent without any required “select a specific trusted item” gate.

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

• Potentially malicious external URL detected (high risk: 0.80). The README explicitly directs installing/running the datadog CLI from https://github.com/leonardocouy/datadog-cli (and the skill runs the CLI via npx), which means remote code from that repository/npm will be fetched and executed as a required runtime dependency.