Microsoft 365 Integration

Access and manage Microsoft 365 email, calendar, OneDrive files, To Do tasks, and contacts via the MS Graph API with authentication support.

This skill appears to implement Microsoft 365 features as described, but proceed cautiously: - Supply-chain risk: the CLI invokes `npx -y @softeria/ms-365-mcp-server` at runtime. That downloads and executes an npm package on demand and is riskier than a pinned release. Prefer installing a specific, vetted package version locally (npm install @softeria/ms-365-mcp-server@<version>) or review the package source on GitHub before using. Consider running the skill in an isolated container/workspace. - Secrets: the skill requires sensitive Azure AD credentials for headless use (client id/secret/tenant). Only provide these if you trust the package and operator. The registry metadata does not declare these env vars — treat that as an inconsistency and ask the publisher to correct it. - Token caching: device-code tokens are cached (location unspecified). Ensure token storage is secure and purge tokens if you remove the skill. - Verify the npm/GitHub project: the SKILL.md attributes the MCP server to Softeria with links; check the upstream project's reputation, examine its code for unexpected network calls, and prefer pinned versions. - Least privilege: when creating the Azure AD app, grant only the minimum Graph permissions you need and avoid broad admin-consent permissions unless required. If you cannot review the upstream npm/github code or are uncomfortable with dynamic npx execution, classify this as too risky to install in a production environment.

Type: OpenClaw Skill Name: ms365 Version: 1.0.0 The skill is classified as suspicious due to its reliance on `npx -y @softeria/ms-365-mcp-server` in `ms365_cli.py` and `README.md`. While this is central to the skill's stated purpose of integrating with Microsoft 365, the `npx -y` command downloads and executes an external Node.js package, introducing a significant supply chain risk. Additionally, the skill requires broad Microsoft Graph API permissions (e.g., `Mail.ReadWrite`, `Files.ReadWrite`, `Chat.ReadWrite`, `Sites.Read.All`) as detailed in `README.md`, which, while necessary for its functionality, represents a high-risk capability if the external dependency were compromised or malicious. There is no direct evidence of intentional malicious behavior within the provided files themselves, nor any prompt injection attempts to subvert the agent for harmful objectives.