Task Router Skill
Task Router
Sync Feishu (Lark) contacts into USER.md so the agent can identify DM senders by name. Use when setting up Feishu identity recognition, updating contacts after HR changes, or configuring multi-user agent access. Feishu DMs only include open_id (no sender name), so this skill embeds an open_id→name lookup table directly in USER.md for zero-tool-call identification.
请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。
SkillHub CLI
skillhub install 4ier/feishu-contacts-sync打开来源仓库,查看说明和兼容环境后,按来源提供的命令或步骤安装。
将来源链接提供给你的 Agent,让它先读取说明、核对依赖,并向你说明将执行的安装步骤。
Feishu (Lark) DM messages only include the sender's open_id — no name. Group messages include Sender metadata with the name, but DMs don't. Without a lookup table, the agent will either:
Embed the full contacts table directly in USER.md. Since workspace files are injected into the system prompt at gateway startup, the agent can match open_id from inbound metadata against the table — zero tool calls needed.
Feishu open_id is scoped to each app. The same person has different open_ids across different Feishu apps. Each OpenClaw instance using a different Feishu app must pull contacts with its own app credentials.
The app needs contact:user.employee_id:readonly or contact:user.base:readonly scope to list users via the contacts API.
python3 scripts/sync_feishu_contacts.py <openclaw_config_path> <feishu_account_name> <user_md_path>
Example:
python3 scripts/sync_feishu_contacts.py ~/.openclaw/openclaw.json my_app ~/workspace/USER.md
Arguments:
openclaw_config_path: Path to your openclaw.json (contains Feishu app credentials)feishu_account_name: The account name under channels.feishu.accounts in your configuser_md_path: Path to your USER.md fileThe script expects USER.md to contain a contacts section with this format:
## 飞书通讯录 (App Name)
飞书 DM 不携带发送者姓名。用 inbound metadata 的 chat_id(格式 `user:ou_xxx`)匹配下表识别发送者。
| 姓名 | open_id |
|------|---------|
| Alice | ou_abc123 |
| Bob | ou_def456 |
On first run, if no contacts section exists, add the section header and description line manually, then run the script to populate the table.
Add t...
这些信息来自收录时的来源分析,可能滞后;使用前请核对最新来源文件。
# Feishu Contacts Sync ## Problem Feishu (Lark) DM messages only include the sender's `open_id` — no name. Group messages include `Sender metadata` with the name, but DMs don't. Without a lookup table, the agent will either: - Assume all DMs are from the person in USER.md (wrong) - Fail to identify the sender entirely ## Solution Embed the full contacts table directly in USER.md. Since workspace files are injected into the system prompt at gateway startup, the agent can match `open_id` from inbound metadata against the table — zero tool calls needed. ## Important: open_id is per-app Feishu `open_id` is scoped to each app. The same person has different open_ids across different Feishu apps. Each OpenClaw instance using a different Feishu app **must pull contacts with its own app credentials**. ## Setup ### 1. Ensure Feishu app has contacts permission The app needs `contact:user.employee_id:readonly` or `contact:user.base:readonly` scope to list users via the contacts API. ### 2. Run the sync script ```bash python3 scripts/sync_feishu_contacts.py <openclaw_config_path> <feishu_account_name> <user_md_path> ``` Example: ```bash python3 scripts/sync_feishu_contacts.py ~/.openclaw/openclaw.json my_app ~/workspace/USER.md ``` Arguments: - `openclaw_config_path`: Path to your openclaw.json (contains Feishu app credentials) - `feishu_account_name`: The account name under `channels.feishu.accounts` in your config - `user_md_path`: Path to your USER.md file ### 3. USER.md format The script expects USER.md to contain a contacts section with this format: ```markdown ## 飞书通讯录 (App Name) 飞书 DM 不携带发送者姓名。用 inbound metadata 的 chat_id(格式 `user:ou_xxx`)匹配下表识别发送者。 | 姓名 | open_id | |------|---------| | Alice | ou_abc123 | | Bob | ou_def456 | ``` On first run, if no contacts section exists, add the section header and description line manually, then run the script to populate the table. ### 4. Add sender identification to AGENTS.md Add t...
Task Router
OpenClaw integration guidance for CAPTCHAS Agent API, including OpenResponses tool schemas and plugin tool registration.
name: modelready description: Start using a local or Hugging Face model instantly, directly from chat. metadata: {"openclaw":{"requires":{"bins": "bash", "curl" }, "env": "URL" }}
Read-only factual data about historical Ethereum mainnet contracts. Use when the user asks about a specific contract address, early Ethereum contracts, deployment era, deployer, bytecode, decompiled code, or do...