The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill implements Microsoft Graph API access through a managed OAuth gateway.
[COMMAND_EXECUTION]: Includes Python examples that use the urllib.request module to interact with the API gateway. These are standard functional examples for an AI agent.
[EXTERNAL_DOWNLOADS]: Communicates with maton.ai, connect.maton.ai, and ctrl.maton.ai for API proxying and session management. These domains are consistent with the vendor's infrastructure as described in the skill metadata.
[DATA_EXFILTRATION]: Authentication is handled securely via the MATON_API_KEY environment variable. Data is only transmitted to the designated service provider for the purpose of fulfilling API requests.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data (emails, calendar events, contacts) which acts as an entry point for potential third-party instructions.
Ingestion points: Email bodies and calendar event descriptions fetched from gateway.maton.ai/outlook/.
Boundary markers: None; external content is passed directly into the agent's context without delimiters.
Capability inventory: Python network requests and file management capabilities inherent to the agent's environment.
Sanitization: No explicit sanitization or filtering of email content is performed before processing.

outlook