← 返回 Skills 市场
Agent Telegram
作者
shangchuanqiytu-ui
· GitHub ↗
· v1.0.0
1460
总下载
0
收藏
13
当前安装
1
版本数
在 OpenClaw 中安装
/install agent-telegram
功能描述
Agent 团队 Telegram 通信规范。所有 Agent 向用户发送消息时必须遵循此规范，确保消息正确发送到 Telegram。
使用说明 (SKILL.md)
Agent Telegram 通信规范
所有 Agent 向用户 (Legend) 发送 Telegram 消息时必须遵循此规范。
账号映射表
| Agent | accountId | Emoji |
|---|---|---|
| main (9527) | default |
🤖 |
| architect (亮亮) | architect |
🏗️ |
| backend (老崔) | backend |
🔧 |
| frontend (小白) | frontend |
🎨 |
| product (小黄) | sproduct |
🟡 |
| content (世龙) | content |
✍️ |
| crawler (湘君) | crawler |
🕷️ |
| qa (赵飞) | qa |
🧪 |
用户 Telegram ID: 5440561025
发送消息
标准格式
message({
action: "send",
channel: "telegram",
accountId: "\x3C你的accountId>",
target: "5440561025",
message: "\x3C你的emoji> \x3C内容>"
})
示例
产品经理发送：
message({
action: "send",
channel: "telegram",
accountId: "sproduct",
target: "5440561025",
message: "🟡 需求文档已完成，请查看：~/Desktop/project/docs/product/001-prd.md"
})
后端工程师发送：
message({
action: "send",
channel: "telegram",
accountId: "backend",
target: "5440561025",
message: "🔧 API 接口开发完成，接口文档：~/Desktop/project/docs/backend/api.md"
})
汇报时机
- ✅ 收到任务时 - 立即汇报"收到任务，开始执行"
- ✅ 每完成子任务 - 汇报完成情况和输出
- ✅ 遇到问题 - 汇报问题并请求决策
- ✅ 任务全部完成 - 汇报最终结果
常见错误
| 错误 | 后果 | 正确做法 |
|---|---|---|
忘记 accountId |
消息发不出去 | 必须指定你的 accountId |
用 sessions_send |
消息不会发到 Telegram | 用 message 工具 |
product 用 accountId: "product" |
账号不存在 | 应该是 sproduct |
| target 写错 | 发给错误的人 | 固定用 5440561025 |
消息模板
任务开始
\x3Cemoji> 收到任务：\x3C任务名>
📝 开始执行...
任务完成
\x3Cemoji> \x3C任务名> 完成
✅ 已完成: \x3C子任务>
📁 输出: \x3C文件路径>
遇到问题
\x3Cemoji> \x3C任务名> 遇到问题
❌ 问题: \x3C描述>
💡 建议: \x3C解决方案或请求决策>
配置位置
- 主配置:
~/.openclaw/openclaw.json→channels.telegram.accounts - 此 Skill:
~/.openclaw/workspace/skills/agent-telegram/SKILL.md
记住：发送 Telegram 消息 = message 工具 + accountId + target: "5440561025"
安全使用建议
This is a usage policy for sending Telegram messages, not executable code — that makes it low-risk from an installation standpoint. However: (1) the spec hardcodes a single Telegram recipient ID (5440561025). If you install and run agents that follow this policy, any content the agent sends (including file paths or attached content) will go to that external account — verify you trust that recipient and that it should receive your data. (2) Examples include explicit local file paths (~/Desktop/...), which can encourage agents to include or read local files; ensure your agent/tooling does not automatically attach or exfiltrate file contents, or remove such examples before use. (3) Check your ~/.openclaw/openclaw.json channels.telegram.accounts mapping before enabling the skill so accountId values align with your own Telegram accounts. If you want to use this skill safely, change the hardcoded target to your own Telegram ID (or make target configurable), remove or sanitize example file paths, and verify how the platform's message tool handles attachments and file reads. If you cannot confirm those details from the skill author or platform docs, exercise caution before installing.
功能分析
Type: OpenClaw Skill
Name: agent-telegram
Version: 1.0.0
The skill bundle defines a communication standard for OpenClaw agents to send messages via Telegram. The `SKILL.md` provides clear instructions for using the `message` tool with a hardcoded target ID, account IDs, and message templates. There is no evidence of prompt injection attempts against the agent, unauthorized data exfiltration, malicious execution, persistence mechanisms, or any other indicators of intentional harmful behavior. The instructions are consistent with a legitimate operational guideline for an AI agent.
能力评估
Purpose & Capability
Name/description match the content: SKILL.md is a communication spec for sending Telegram messages. It does not request unrelated binaries, env vars, or installs, which is proportionate to a policy/instruction-only skill.
Instruction Scope
The instructions require all agents to send messages to a single hardcoded Telegram target (target: "5440561025") and specify concrete local file paths in examples. While sending to Telegram is the stated goal, the fixed external recipient and example content that references local files increase the risk that agents will unwittingly transmit sensitive local data. The spec also references the local config path (~/.openclaw/openclaw.json) which is reasonable to document but could encourage agents to read or rely on local configuration; the doc does not explicitly limit what parts of the filesystem may be attached to messages.
Install Mechanism
Instruction-only skill with no install steps and no code files; lowest install risk.
Credentials
No environment variables, credentials, or config paths are required by the skill beyond documenting where OpenClaw's Telegram accounts live. Requested surface is minimal and consistent with the described purpose.
Persistence & Privilege
always is false and the skill is user-invocable. It does not request permanent inclusion or modify other skills' configs. No elevated persistence privileges are requested.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install agent-telegram - 安装完成后，直接呼叫该 Skill 的名称或使用
/agent-telegram触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
- Initial release of the agent-telegram skill.
- Defines the communication protocol for all Agents when sending messages to users via Telegram.
- Provides a comprehensive account mapping table for Agents, including accountId and associated emoji.
- Specifies standard message format and templates for various reporting scenarios.
- Lists common errors and correct usage for successful Telegram delivery.
- Outlines configuration file locations for both the main settings and this skill.
元数据
常见问题
Agent Telegram 是什么？
Agent 团队 Telegram 通信规范。所有 Agent 向用户发送消息时必须遵循此规范，确保消息正确发送到 Telegram。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 1460 次。
如何安装 Agent Telegram？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install agent-telegram」即可一键安装，无需额外配置。
Agent Telegram 是免费的吗？
是的，Agent Telegram 完全免费（开源免费），可自由下载、安装和使用。
Agent Telegram 支持哪些平台？
Agent Telegram 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Agent Telegram？
由 shangchuanqiytu-ui（@shangchuanqiytu-ui）开发并维护，当前版本 v1.0.0。
推荐 Skills