REA
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
CLI for Telegram — agent-friendly, daemon-based, with webhook event push.
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/skills/tlgrcli-tlgr)在同一工作中把 AI 连到服务的 MCP
Reverse engineer anything from your terminal or agent with one CLI and MCP server.
Codebase intelligence for TypeScript and JavaScript, exposed as structured MCP analysis tools.
MCP server for terminal commands, file operations, and process management
同类工作,排序与 Skill 列表相同
| 本项 tlgr | rea | andrej-karpathy-skills | graphify | MoneyPrinterTurbo | |
|---|---|---|---|---|---|
| 作者 | tlgrcli | morluto | multica-ai | Graphify-Labs | harry0703 |
| Skill 数 | 1 | 1 | 1 | 1 | 1 |
| GitHub 星标 | 201 | 463,670 | 233,682 | 148,887 | 143,987 |
| 最近代码更新 | 2026年10月28日 | 2026年11月17日 | 2026年4月20日 | 2026年11月17日 | 2026年11月16日 |
| 许可证 | MIT | MIT | – | Apache-2.0 | MIT |
tlgrcli 提供的 1 个 Skill(指令集),AI 读取后照做 · 不是 MCP server
MCP = 手
把 AI 连到外部服务或数据,例如读取 Notion 或开发票。需要有程序或服务器在运行。
Skill = 手册
教 AI 怎么做某项工作的文本文件,不连接任何服务。
有些会重叠:少数 Skill 带有调用工具的脚本,少数 MCP 只提供知识。请从你想完成的工作出发。
Claude CodeCodexCursor
依据开发者 README 和 repo 中的文件夹
GitHub account tlgrcli (organization) is only 8 months old: take extra care
★ 201 stars, few users so far · code last changed 2026年10月28日
No, plain text files the AI reads and follows
No scripts, so it does not go online by itself
No hooks or writes to machine settings found
Checked automatically from the repo files at commit 9b441f2 (read only, nothing was run). Not a safety guarantee: read it and decide for yourself
第一步:你用哪个 AI?
选择你用的应用,本页只显示该应用的安装方法
Uploading to claude.ai would fail: the description is over 200 characters (441). Use "Let the AI install it" instead
复制这句话,粘贴到 Claude Code 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
重新打开 Claude Code,输入 /tlgr,或直接提出相关任务,Claude 会自动选用
Checked against official docs 2026年11月12日 · Source: code.claude.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
已经装好了吗?先检查 · 打开终端,输入这条命令后按 Enter
node -vMac:按 ⌘ + Space,输入 Terminal 后按 Enter
Windows:打开“开始”菜单,输入 Windows PowerShell,然后选“打开”
还没安装: 打开官方下载页,选择 LTS 版本和点选式安装程序:Mac 用 .pkg · Windows 用 .msi。双击后一路下一步即可,不需要 curl 命令
装好后关掉终端再打开,重新运行同一条命令,应显示版本号: node -v
安装一次,所有需要 Node.js 的工具都能用
正在检查命令 · 通过后才会显示复制按钮
摘自开发者 README,会下载并运行开发者的代码
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
ChatGPT 桌面应用内置 Codex,Codex 从 ~/.agents/skills 文件夹读取 Skill
在 ChatGPT 桌面应用中打开 Codex,把这句话粘贴到对话中
The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
把 tlgr 文件夹放到 ~/.agents/skills/tlgr
~ 是你的主文件夹($HOME)
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
不能使用你自己安装的 Skill,只能使用目录中插件自带的 Skill
想在 ChatGPT 中使用这个 Skill,请按上面的方法使用 ChatGPT 桌面应用
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
复制这句话,粘贴到 Codex 的对话中 · 适用于 ChatGPT 桌面应用中的 Codex、Codex CLI 和 IDE 扩展
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Codex 会自动发现新安装的 Skill;如果没有出现,请重启 Codex · 使用方法:在 ChatGPT 中输入 @ 并选择 Skill · 在 Codex CLI 或 IDE 扩展中运行 /skills,或输入 $ 加 Skill 名称
Checked against official docs 2026年11月15日 · Source: learn.chatgpt.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
复制这句话,粘贴到 Cursor 的对话中
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
在 Cursor 的 Agent 聊天中输入 / 并选择 tlgr
Checked against official docs 2026年11月12日 · Source: cursor.com
Skill 是 AI 会照做的指令。安装前先阅读文件,只安装可信开发者的 Skill。
Mac:按 Command+空格输入 Terminal · Windows:下方命令在 PowerShell 中无法使用,请改用上方的“让 AI 帮你安装”
这条命令把 repo 下载到临时文件夹,只复制含 SKILL.md 的文件夹
正在检查命令 · 通过后才会显示复制按钮
适用于其他能读取 Skill 的应用。这句话列出了 Claude Code、Codex 和 Cursor 的路径,AI 会按你的应用选择 · Copy this sentence into the app’s chat
For one project, paste it into a chat that is open in that project’s folder · The AI always shows you the path before installing
The AI tells you what it will do first and asks your permission at every step
Checked against official docs 2026年11月12日 · Source: code.claude.com, learn.chatgpt.com, cursor.com
其他支持 Agent Skills 的应用(如 Gemini CLI、GitHub Copilot)使用不同文件夹,请查看该应用的指南,再把含 SKILL.md 的文件夹复制过去。
已于 2026年11月11日 对照 Claude Code、Codex 和 Cursor 官方指南核实 skill 文件夹
以下文字来自开发者的 repo,不是 PopNx MCP 的指令
6,162 个字符 · 在 GitHub 打开完整文件
---
name: tlgr
description: Read and act on a personal Telegram account from the terminal with the tlgr CLI (MTProto user account, not a bot). Use when the user wants to check unread Telegram chats, read or search message history, send, reply to, edit, forward or schedule messages, manage chats, groups, channels, contacts, folders, reactions, polls, media or stories, or receive Telegram events through a webhook. Every command answers in JSON with stable exit codes.
metadata: {"openclaw": {"requires": {"bins": ["tlgr"]}, "os": ["darwin", "linux"]}}
---
# tlgr: Telegram for agents
`tlgr` controls a real Telegram user account over MTProto through a local
daemon. Everything you can do in a Telegram app, you can do from a shell:
messages, chats, groups, channels, contacts, media, stories, polls and more.
The full reference is `AGENT.md` in the repository, and `tlgr schema --json`
describes every command, flag and response.
## Install and check
```bash
pipx install tlgr-cli # the command is `tlgr`
tlgr agent whoami --json # active account, daemon health, schema version
tlgr status --check # non-zero exit when anything is wrong
```
The daemon starts itself on first use. If `whoami` reports no account, go to
**Logging in**.
## Rules that matter
1. **Always pass `--json`.** Errors are JSON on stdout too, and the exit code
is the contract. `--results-only` strips the envelope and `--select a,b.c`
projects fields.
2. **Reading can be visible.** `chat open` sends a read receipt the other side
sees and clears the owner's unread badge. To look without a trace use
`chat catchup`, `message list` or `chat open --no-read`. A read receipt
cannot be taken back; `chat unread` only restores the owner's badge.
3. **Name the account when there is more than one.** Pass `-a <alias>` so a
message never goes out from the wrong identity.
4. **Secrets never go in argv.** Use `--password-env`, `--api-hash-env`,
`--token-env` and friends, or their `-stdin` and `-file` forms.
5. **Chats are `@username` or a numeric id.** Display names and phone numbers
are not accepted. Put negative ids after `--`:
`tlgr message list --json --limit 5 -- -100123456`.
6. **Preview before destructive calls** with `--dry-run`, and confirm with the
user before deleting, leaving, blocking or sending to someone new.
7. **Respect rate limits.** Exit 7 carries `wait_seconds`; wait that long
before retrying. Exit 13 means the answer is unknown, never treat it as a
"no".
## Everyday loop
```bash
tlgr catchup --json # every unread chat with recent messages, no read receipts
tlgr message list @alice --json --limit 20 # silent history
tlgr message search @team "invoice" --json # search one chat
tlgr message send @alice "On my way" --json # send
tlgr message send @alice "Yes" --reply-to 4211 --json
tlgr message edit @alice 4212 "Yes, 5pm" --json
tlgr message forward @news 88 90 --to @archive --json
tlgr chat open @alice --json # history plus a visible read receipt
```
Long text over 4096 UTF-16 units is refused unless you add `--split`.
`--typing-auto` shows "typing..." for a length-based duration before sending.
`--schedule TS` schedules instead of sending now.
Lists return `{items, has_more, next_cursor}`. Pass `--cursor <next_cursor>`
until `has_more` is false, or use `--all`.
## Beyond messages
Each group has `--help` and a generated page in `docs/reference/`:
| Group | Examples |
|---|---|
| `chat` | `list --unread`, `get --full`, `posters`, `members`, `archive`, `mute`, `leave` |
| `contact`, `user`, `resolve` | find, add and inspect people; turn a link or username into a peer |
| `reaction`, `poll`, `todo`, `location` | react, vote, create polls and checklists, share places |
| `media`, `sticker`, `gif`, `emoji` | upload, download, export |
| `story`, `folder`, `profile`, `privacy` | post stories, organise folders, change settings |
| `account` | `list`, `check`, `session list --unconfirmed` |
Before planning something unusual, run `tlgr agent capabilities --json`. It
separates what this build cannot do, what this account may not do (Premium,
bot-only, admin-only) and what tlgr refuses on purpose, each with a reason.
## Logging in
Login is two ordinary commands, so only reading the code needs a person:
```bash
tlgr auth send-code +15551234567 --alias main --api-id 12345 --api-hash-env TLGR_API_HASH --json
tlgr auth verify-code 12345 --alias main --json
```
Ask the user for the code Telegram sends them. Exit 4 with
`AUTH_PASSWORD_REQUIRED` means two-step verification is on: rerun
`verify-code` with `--password-env TLGR_2FA_PASSWORD` after the user sets
that variable. Never ask the user to paste a password into the chat.
## Reacting to events
To be woken by Telegram instead of polling, the daemon can push events to an
HTTP endpoint, signed with HMAC:
```bash
tlgr webhook set --url https://example.com/hooks/agent \
--events message_new,message_edited --secret-env TLGR_WEBHOOK_SECRET --enabled
tlgr webhook test
```
`tlgr events list --json` lists all event types. Deterministic auto-replies and
forwards that need no model can run in the daemon as jobs in
`~/.tlgr/jobs.yaml`.
## Limiting what you can do
`--enable-commands message.list,message.send,chat.list` (or
`TLGR_ENABLE_COMMANDS`) allows only those operations; anything else exits 6.
Use it when the user wants you read-only or restricted to a few actions. It is
a guard rail, not a security boundary.
## Exit codes
| Code | Meaning | What to do |
|---|---|---|
| 0 | success | `meta.already: true` means it was already done |
| 2 | bad arguments | fix the call; check `tlgr schema <command> --json` |
| 3 | empty | nothing matched |
| 4 | auth needed | log in again |
| 5 | not found | wrong chat or id |
| 6 | permission denied or blocked by `--enable-commands` | do not retry |
| 7 | rate limited | wait `wait_seconds` |
| 8 | transient | retry shortly |
| 11, 12 | daemon or IPC problem | `tlgr daemon status --json`, then `tlgr daemon restart` |
| 13 | indeterminate | report as unknown |
提问、分享技巧或报告无法使用。按“有帮助”数量排序。