← 返回 Skills 市场
Telegram CS Agent
作者
youziyouzishu
· GitHub ↗
· v1.0.1
· MIT-0
98
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install tg-cs-agent
功能描述
Deploy and manage a Telegram customer service bot powered by Claude + RAG. Use when setting up a new CS bot, adding knowledge base docs, managing the bot lif...
安全使用建议
This package appears to implement a Telethon-based Telegram userbot that calls Anthropic and stores a local ChromaDB vector store — that's consistent with a RAG customer-service bot. Before installing:
- Confirm the registry metadata mismatch: the skill requires ANTHROPIC_API_KEY and TELEGRAM_API_ID/TELEGRAM_API_HASH (and will prompt for your phone/2FA); the registry listing incorrectly shows no required env vars.
- Understand 'userbot' vs 'bot': this uses a user account (Telethon) and will create a persistent session at ~/.tgctl-telethon. A user account may access more chats and actions than a Bot API token. Only use an account you trust to attach to this code.
- Run in an isolated environment (dedicated account or VM/container) and review the .env you create; Anthropic use incurs costs tied to your API key.
- Review and control what markdown/docs you load: docs are vectorized and stored locally in ChromaDB; do not load sensitive data unless you accept local persistence.
- If you need a less privileged integration, consider adapting the code to use a bot token (Bot API) instead of a user session, or require an organization review of the skill before deployment.
If you want, I can list exact lines that read/env or file paths the skill will create and where sensitive data will be stored.
功能分析
Type: OpenClaw Skill
Name: tg-cs-agent
Version: 1.0.1
The skill bundle provides a legitimate implementation of a RAG-powered Telegram customer service bot using Claude, Telethon, and ChromaDB. The code follows standard patterns for such applications, including environment variable configuration, vector database management in `knowledge.py`, and a Telethon-based client in `telegram_client.py`. There is no evidence of data exfiltration, malicious execution, or prompt injection; the 'handoff' mechanism and Telegram session handling are transparently documented and aligned with the stated purpose of managing a customer service bot for the 'Seers Markets' platform.
能力评估
Purpose & Capability
The code, SKILL.md, and included dependencies match the stated purpose: it uses Anthropic, Telethon (user account), ChromaDB and embeddings to provide a RAG-powered Telegram CS agent. However the registry metadata claims 'Required env vars: none' while the SKILL.md and code clearly require ANTHROPIC_API_KEY, TELEGRAM_API_ID, TELEGRAM_API_HASH, and related settings — an inconsistency you should verify with the publisher.
Instruction Scope
Runtime instructions and code stay within the described scope: they load local markdown docs, embed them, query Anthropic, and relay replies via Telethon. Important operational behaviors are explicit but significant: this is a Telethon userbot (not the Bot API), it will prompt for a phone number/2FA during interactive login, and it stores a persistent Telethon session under ~/.tgctl-telethon. The skill sends user messages to Anthropic (the intended LLM) and not to any other network endpoints.
Install Mechanism
No packaged install spec; SKILL.md instructs pip installing standard packages (anthropic, chromadb, sentence-transformers, telethon, python-dotenv). There are no external download URLs or obfuscated installers. The requirements are proportional to the functionality (embedding models will be downloaded by sentence-transformers at runtime).
Credentials
The environment/credentials required by the code (Anthropic API key, Telegram API ID/HASH, optional HANDOFF_CHAT_ID) are appropriate for this skill, but the registry metadata omitted declaring them. Also note the skill will prompt for and store your Telegram account session (phone number + potential 2FA) — handing a user account session to the code grants broader permissions than a bot-token-based integration.
Persistence & Privilege
The skill does not request elevated platform privilege (always:false). It does create persistent artifacts: Telethon session files under ~/.tgctl-telethon and a local ChromaDB collection (vector DB). These are normal for this use case but mean data (session, vectorized docs) is stored on disk and should be managed/isolated if you have security/privacy concerns.
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 tg-cs-agent - 安装完成后，直接呼叫该 Skill 的名称或使用
/tg-cs-agent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
v1.0.1
Auto-login: interactive Telegram auth when no session exists, no need for tgctl-telethon
v1.0.0
Initial release: Claude + RAG + Telethon customer service bot with multilingual support
元数据
常见问题
Telegram CS Agent 是什么？
Deploy and manage a Telegram customer service bot powered by Claude + RAG. Use when setting up a new CS bot, adding knowledge base docs, managing the bot lif... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 98 次。
如何安装 Telegram CS Agent？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install tg-cs-agent」即可一键安装，无需额外配置。
Telegram CS Agent 是免费的吗？
是的，Telegram CS Agent 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Telegram CS Agent 支持哪些平台？
Telegram CS Agent 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 Telegram CS Agent？
由 youziyouzishu（@youziyouzishu）开发并维护，当前版本 v1.0.1。
推荐 Skills