operations-team

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow uses MCP tools for Notion and Slack, which can ingest outsider-authored free text (e.g., Notion page contents or Slack messages/comments from other users) into the agent’s LLM context via those tool calls.

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

• Potentially malicious external URL detected (high risk: 0.90). The Required Tools declare runtime npx commands that fetch-and-execute remote npm packages the skill depends on (npx -y @makenotion/mcp-server and npx -y @modelcontextprotocol/server-slack), so external code would be retrieved and run at runtime.