多机器人去重消息

多飞书机器人去重回复规则。当同一用户通过多个机器人渠道同时发送相同消息时，AI 只回复一次，避免重复打扰。适用于配置了多路飞书机器人接入的 OpenClaw 场景。

This skill is internally consistent and simple, but review these before installing: - Confirm you trust the source (homepage unknown). The skill will store sender IDs, timestamps, and message hashes in skills/multi-bot-dedup/dedup_state.json. - Decide on retention and privacy: remove or rotate entries if you don't want persistent PII. Consider hashing with a salt or storing only non-reversible identifiers if privacy is a concern. - Ensure the runtime has safe file permissions for the skills directory so other users/processes cannot read the state file if that would expose PII. - The SKILL.md leaves hashing and update atomicity unspecified — in multi-process or multi-instance setups you may get race conditions; consider using a central atomic store (Redis/DB) if you need reliable dedup across concurrent workers. - Test in a staging environment to confirm the 30s window and first-50-characters rule meet your needs and don't cause unexpected suppression of legitimate replies.

Type: OpenClaw Skill Name: multi-bot-dedup Version: 1.0.0 The skill bundle provides behavioral instructions for an AI agent to implement a deduplication mechanism for multiple Feishu (Lark) bots. It uses a local state file (dedup_state.json) to track message hashes and timestamps, ensuring the agent only responds once to identical messages within a 30-second window. No malicious code, data exfiltration, or harmful prompt injections were identified.