提供中英文邮件草稿生成、回复建议、模板应用及收件箱邮件分类整理功能，支持正式与非正式语气调整。

This skill appears to be a local email template/draft generator and does not actually connect to your mailbox. Before installing: - Be aware the skill will create and update JSON files (templates.json and sent_emails.json) in the skill's directory (the code uses the skill folder; SKILL.md mentions ~/.openclaw/... — confirm where files will end up on your system). - The description/SKILL.md claim inbox classification and '整理收件箱', but the code does not access any mailboxes or perform real inbox organization. If a future version requests email credentials or network access to integrate with your mail client, treat that as a sensitive escalation and review carefully. - No credentials or network access are requested now, which limits remote exfiltration risk, but generated emails and any sensitive content you paste into the skill will be stored locally—consider whether that is acceptable for your workflow. - If you want actual client integration/automatic replies, expect that additional permissions (SMTP/IMAP/OAuth tokens) would be required; only grant those if you trust the publisher and have reviewed the code. - If you proceed, consider running the skill in a restricted/sandboxed environment or examining/adjusting the DATA_DIR path in the Python file to point to a location you control (e.g., a private workspace) to avoid unexpected file placement. Overall: the implementation is coherent for local template generation but inconsistent with some advertised capabilities; this looks more like incomplete or sloppy documentation than malicious behavior, but exercise caution and validate file storage location and future permission requests.

Type: OpenClaw Skill Name: email-helper Version: 1.0.0 The Email Helper skill is a legitimate utility for generating email drafts and templates based on user queries. The Python script (email_helper.py) implements template-based generation, tone adjustment, and local statistics tracking using standard libraries without any network activity, unauthorized file access, or code execution vulnerabilities. The instructions in SKILL.md and README.md are consistent with the provided code logic.