报告双通道智能推送

将分析报告智能摘要推送到微信(≤100字任务清单)+全文推送到QQ邮箱,说'推送报告'时触发

This skill mostly does what it says (send mail via QQ SMTP and post to a WeChat push endpoint) and only requires the expected environment variables. Before installing or enabling it: 1) fix or confirm the script path (SKILL.md calls scripts/push_report.py but the code file is scripts/ush_report.py) — otherwise the agent will fail to run the sender; 2) don't assume the promised "built-in" compliance, rate-limiting, or deduplication actually exists — the shipping script does not implement these protections, so ensure the agent enforces them before calling the script or add those checks to the script; 3) verify which WeChat push service you will use — the script always uses PushPlus endpoint, so Server酱-style tokens (SCT...) will not work; 4) note the script uses HTTP for pushplus (consider changing to HTTPS if supported) and truncates summaries locally; 5) test with non-sensitive content and test accounts first (so your real auth codes are not exposed during debug). These inconsistencies explain the "suspicious" rating; they may be benign packaging errors but should be resolved before trusting the skill with real credentials or sensitive reports.

Type: OpenClaw Skill Name: report-push-wechat-qqmail Version: 1.0.0 The skill bundle is designed to push report summaries to WeChat (via the PushPlus service) and full reports to QQ Mail. The implementation in `scripts/push_report.py` uses standard libraries (smtplib and requests) to perform these actions using user-configured environment variables. The instructions in `skill.md` include safety checks for content compliance and do not contain any malicious prompt injection or unauthorized data exfiltration logic.