datadog-automation

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

• Third-party content exposure detected (medium risk: 0.30). Runtime workflow includes DATADOG_SEARCH_LOGS / log-search queries in “Search and Analyze Logs,” which can ingest outsider-authored log text that originates from users/app traffic and may contain prompt-injection strings.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill requires connecting to the external MCP endpoint https://rube.app/mcp at runtime (see SKILL.md) and instructs the agent to call RUBE_SEARCH_TOOLS to fetch tool schemas, meaning remote content from that URL can control agent prompts/tooling, so it is a runtime dependency that can influence agent instructions.