feishu-message-group-whitelist
v1.0.0飞书群聊白名单关键词过滤。当飞书机器人收到群消息时，根据白名单判断是否需要回复——只有消息中包含白名单关键词时才触发回复，实现精准触发而非广播所有消息。触发词示例：群里有人发消息时、提到特定关键词时、需要过滤无关消息时。
⭐ 0· 72·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (Feishu group whitelist keyword filter) aligns with the provided instructions and files: it only needs to read a local config/whitelist.txt and decide whether to invoke the normal agent response flow. No unrelated credentials, binaries, or config paths are requested.
Instruction Scope
SKILL.md is narrowly scoped: it instructs the agent to read config/whitelist.txt (skill root) and only trigger the normal response when a keyword matches. One operational caveat: if the whitelist is empty the instructions explicitly say to 'skip filtering' and respond to all messages — that can cause broad replies and increased token usage if the operator forgets to populate the file.
Install Mechanism
No install spec or code is included (instruction-only skill). Nothing will be downloaded or written by an installer as part of skill setup.
Credentials
The skill requests no environment variables, credentials, or external configuration. This is proportionate for a local whitelist filter.
Persistence & Privilege
always:false (default); model invocation is allowed (normal). The skill does not request permanent global presence or modify other skills or system-wide settings.
Assessment
This skill appears to do exactly what it says: read config/whitelist.txt and only forward messages that contain a listed keyword. Before installing or enabling it, consider the following: (1) Populate config/whitelist.txt with the intended trigger words — if left empty the skill will allow replies to all messages, which may cause unwanted broadcasting and extra token/model usage. (2) Confirm you trust the skill source since the skill will pass matched message content into your agent's normal response flow (the agent may send that content to the model or external systems depending on your agent configuration). (3) Verify the skill's config directory is not symlinked to sensitive files you don't want the skill to read. (4) Test in a controlled group or sandbox to ensure behavior matches expectations.Like a lobster shell, security has layers — review code before you run it.
latestvk976xs1987rzsas2crmh51pvn183s532
License
MIT-0
Free to use, modify, and redistribute. No attribution required.