Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

mmxagent-skill-feishu

v1.0.0

连接飞书机器人。用接口直连替换 `feisu.md` 的第 1 步,拿到链接(必须拼接 from=maxclaw)并在用户完成配置后 poll 出 `client_id/client_secret`,再继续走 `feisu.md` 的第 2 步和第 3 步。适用于连接飞书、接入飞书、绑定飞书机器人、创建飞书机器人...

0· 76·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for oreoandyuumi/mmxagent-skill-feishu.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "mmxagent-skill-feishu" (oreoandyuumi/mmxagent-skill-feishu) from ClawHub.
Skill page: https://clawhub.ai/oreoandyuumi/mmxagent-skill-feishu
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install mmxagent-skill-feishu

ClawHub CLI

Package manager switcher

npx clawhub@latest install mmxagent-skill-feishu
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The name/description say this replaces step 1 of an existing feisu.md flow to obtain an OAuth link and then poll for client_id/client_secret. The SKILL.md only uses Feishu endpoints (accounts.feishu.cn, open.feishu.cn) and curl — these are consistent with the stated purpose. Note: the skill mandates appending &from=maxclaw to the verification link (a tracking/origin parameter) which is not explained in the description and may tie registrations to the skill author/operator.
!
Instruction Scope
Instructions directly call Feishu endpoints via curl, create and use a cookie jar in /tmp, extract verification_uri_complete/device_code/user_code, then poll to receive client_id and client_secret. That flow is coherent but sensitive: the skill will surface client_secret and instruct the agent/user to handle it. The SKILL.md also instructs the agent to explicitly avoid explaining or executing admin/permission/event-subscription steps — this could omit necessary security/approval steps for some orgs and may be attempting to bypass normal administrative controls.
Install Mechanism
Instruction-only skill with no install spec, no downloaded code or third-party packages. Lowest installation risk.
Credentials
The skill requests no environment variables or external credentials up front (good). However, its runtime instructions produce and reveal sensitive credentials (client_id/client_secret) obtained from Feishu; users should treat those secrets carefully. There are no unrelated credential requests.
Persistence & Privilege
The skill is not always:true and does not request persistent system-level privileges. It writes a transient cookie jar to /tmp, which is reasonable for this flow.
What to consider before installing
This skill performs a direct API registration flow with Feishu (curl to accounts.feishu.cn), obtains client_id and client_secret, and requires you to complete the browser step — it will display the client_secret in cleartext. Before using it: verify you trust the skill author (the enforced &from=maxclaw parameter may tag registrations), confirm that bypassing the official installer is acceptable for your organization, and ensure you do not paste client_secret into untrusted places. If your org requires admin approvals or specific permission steps, prefer the official documented flow or consult your Feishu admin; do not use this skill if you need the agent to follow those admin processes automatically.

Like a lobster shell, security has layers — review code before you run it.

latestvk971x5d22qb1zb5xp1eyjze5b9843mss
76downloads
0stars
1versions
Updated 3w ago
v1.0.0
MIT-0

Feishu Connect Skill

适用场景

用户要求连接飞书、绑定飞书机器人、创建飞书机器人、OpenClaw 接飞书时触发。

最高优先级规则(必须遵守,违反即失败)

  1. 给用户的链接必须拼接 &from=maxclaw,没有这个参数的链接是错误的。
  2. 这个 skill 只替换 feisu.md 的第 1 步;用户回复"ok"或"配置好了"之后,不要再重复执行 feisu.md 的第 1 步,直接继续走 feisu.md 的第 2 步和第 3 步。
  3. 用户回复"ok"或"配置好了"之后,禁止输出任何关于"配置权限"、"开启机器人能力"、"配置事件订阅"、"发布应用"、"管理员审批"、"权限管理"、"批量导入"的内容。

核心原则

  • 直接走接口直连模式,不要先跑官方安装器(npx -y @larksuite/openclaw-lark install)。
  • 这个 skill 仅用于替换 feisu.md 的第 1 步;拿到凭证后,后续配置和配对继续按 feisu.md 的第 2 步和第 3 步执行。
  • 链接必须裸露返回,不要用反引号、代码块包裹,用户必须能直接点击。
  • 不要自动开轮询,给链接后等用户确认再继续。
  • 不要生成二维码图片。
  • 不要手动卸载/禁用/删除插件,不要手动改写配置文件。

执行流程

第 1 步:接口直连,一次拿到链接

COOKIE_JAR=/tmp/feishu_cookies.txt
rm -f "$COOKIE_JAR"

# init
INIT_RESP=$(curl -s -c "$COOKIE_JAR" -b "$COOKIE_JAR" \
  -X POST "https://accounts.feishu.cn/oauth/v1/app/registration" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "action=init")
echo "=== INIT ==="
echo "$INIT_RESP"

# begin
BEGIN_RESP=$(curl -s -c "$COOKIE_JAR" -b "$COOKIE_JAR" \
  -X POST "https://accounts.feishu.cn/oauth/v1/app/registration" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "action=begin&archetype=PersonalAgent&auth_method=client_secret&request_user_info=open_id")
echo "=== BEGIN ==="
echo "$BEGIN_RESP"

从 begin 返回值中提取:

  • verification_uri_complete — 原始链接
  • device_code — 保存下来,用户确认后 poll 用
  • user_code — 从链接参数中提取

第 2 步:拼接 from=maxclaw,生成最终链接

拿到 verification_uri_complete 后,必须在末尾拼接 &from=maxclaw

示例:

没有 &from=maxclaw 的链接是错误的,不要发给用户。

第 3 步:把链接直接发给用户

链接裸露返回。正确输出:


请在浏览器中直接打开这个链接完成飞书配置:

https://open.feishu.cn/page/openclaw?user_code=XXXX-XXXX&from=maxclaw

用户码:XXXX-XXXX

配置完成之后回来和我说一声"ok"或者"配置好了"就行。


禁止把链接放在反引号或代码块里。

第 4 步:用户回复"ok"或"配置好了"之后

先执行一次 poll:

curl -s -c "$COOKIE_JAR" -b "$COOKIE_JAR" \
  -X POST "https://accounts.feishu.cn/oauth/v1/app/registration" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d "action=poll&device_code=<之前保存的 device_code>"

返回 client_id + client_secret → 成功。 返回 authorization_pending → 告诉用户飞书侧似乎还没完成,让用户再确认。

poll 成功后,直接告诉用户: "配对完成了!现在请在飞书里找到你的机器人,给它发一条消息,机器人会回复一个配对码。把这个码告诉我,我来帮你完成配对授权。"


说明这里已经完成了 feisu.md 的第 1 步替换,并拿到了后续要用的凭证:

  • client_id 作为 App ID
  • client_secret 作为 App Secret

然后不要回到 feisu.md 的第 1 步,直接继续执行:

  • feisu.md 的第 2 步:Configure OpenClaw
  • feisu.md 的第 3 步:Start + test

也就是说,后面继续按 feisu.md 去配置 OpenClaw、启动 gateway、给机器人发消息,并用配对码执行 openclaw pairing approve feishu <CODE>


不要再输出飞书开放平台里的建应用、配权限、开事件订阅、发布应用这类第 1 步指引。

禁止事项

  • 给用户的链接不拼 &from=maxclaw
  • 在用户确认后又重新执行 feisu.md 的第 1 步
  • 输出"配置权限/开启机器人能力/配置事件订阅/发布应用/管理员审批/批量导入/权限JSON"等内容
  • 先跑官方安装器再切直连
  • 手动卸载/禁用插件
  • 自动开轮询
  • 生成二维码图片
  • 用反引号/代码块包裹链接
  • authorization_pending 当失败

一句话总结

用这个 skill 替换 feisu.md 的第 1 步:curl 拿链接 → 拼 &from=maxclaw → 裸链接给用户 → 等用户说 ok → poll 拿到 client_id/client_secret → 继续走 feisu.md 的第 2 步和第 3 步

Comments

Loading comments...