The Agent Skills Directory

[SAFE]: The skill uses the dws CLI for DingTalk event monitoring. No malicious commands, remote code execution, or persistence patterns were found in the instructions or references.\n- [SAFE]: Sensitive data protection is implemented for VoIP events; the skill documentation mandates masking the room_code by default to prevent exposure in logs or to the AI Agent context.\n- [SAFE]: Secure configuration management is observed, with instructions defining restricted file permissions (0600 for files, 0700 for directories) for local subscription state and identity-hashed configuration paths.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external message data but mitigates potential injection attacks by requiring the use of stable, system-provided identifiers (conversation_id, open_dingtalk_id) for all downstream actions, rather than relying on untrusted user-controlled display names.

dingtalk-event