telegram-reader（Telegram reader 搜索）

Telegram News Skill (Read-Only)

Reads Telegram channels and groups for financial news and market research using tdl, a Telegram CLI tool.

This skill is read-only. It is designed for financial research: reading channel messages, monitoring financial news channels, and exporting message history. It does NOT support sending messages, joining/leaving channels, or any write operations.

Step 1: Ensure tdl Is Installed

Current environment status:

!`(command -v tdl && tdl version 2>&1 | head -3 || echo "TDL_NOT_INSTALLED") 2>/dev/null`

If the status above shows a version number, tdl is installed — skip to Step 2.

If TDL_NOT_INSTALLED, install tdl based on the user's platform:

Ask the user which installation method they prefer. Default to Homebrew on macOS, curl script on Linux.

Step 2: Ensure tdl Is Authenticated

Current auth status:

!`(tdl chat ls --limit 1 2>&1 >/dev/null && echo "AUTH_OK" || echo "AUTH_NEEDED") 2>/dev/null`

If AUTH_OK, skip to Step 3.

If AUTH_NEEDED, guide the user through login. Login requires interactive input — the user must enter their phone number and verification code manually.

Login methods

Method A: QR Code (recommended — fastest)

tdl login -T qr

A QR code will be displayed in the terminal. The user scans it with their Telegram mobile app (Settings > Devices > Link Desktop Device).

Method B: Phone + Code

tdl login -T code

The user enters their phone number, then the verification code sent to their Telegram app.

Method C: Import from Telegram Desktop

If the user has Telegram Desktop installed and logged in:

tdl login

This imports the session from the existing desktop client. The desktop client must be from the official website, NOT from the App Store or Microsoft Store.

Namespaces

By default, tdl uses a default namespace. To manage multiple accounts:

tdl login -n work -T qr      # Login to "work" namespace
tdl chat ls -n work           # Use "work" namespace for commands

Important login notes

• Login is a one-time operation. The session persists on disk after successful login.
• If login fails, ask the user to check their internet connection and try again.
• Never ask for or handle Telegram passwords/2FA codes programmatically — always let the user enter them interactively.

Step 3: Identify What the User Needs

Match the user's request to one of the read operations below.

Chat identifiers

The -c flag accepts multiple formats:

Step 4: Execute the Command

Listing chats

# List all chats
tdl chat ls

# JSON output for processing
tdl chat ls -o json

# Filter for channels only
tdl chat ls -f "Type contains 'channel'"

# Search by name
tdl chat ls -f "VisibleName contains 'Bloomberg'"

Exporting messages

Always use --all --with-content to get text messages (not just media):

# Last 20 messages from a channel
tdl chat export -c @channel_name -T last -i 20 --all --with-content -o /tmp/tdl-export.json

# Messages from a time range (Unix timestamps)
tdl chat export -c @channel_name -T time -i 1710288000,1710374400 --all --with-content -o /tmp/tdl-export.json

# Messages by ID range
tdl chat export -c @channel_name -T id -i 100,200 --all --with-content -o /tmp/tdl-export.json

Key rules

1. Check auth first — run tdl chat ls --limit 1 before other commands to verify the session is valid
2. Always use --all --with-content when exporting messages for reading — without these flags, tdl only exports media messages
3. Use -o FILE to save exports to a file, then read the JSON — this is more reliable than parsing stdout
4. Start with small exports — use -T last -i 20 unless the user asks for more
5. Use filters on chat ls to help users find the right channel before exporting
6. NEVER execute write operations — this skill is read-only; do not send messages, join channels, or modify anything
7. Convert timestamps — when the user gives dates, convert to Unix timestamps for the -T time filter

Working with exported JSON

After exporting, read the JSON file and extract the relevant information:

# Export messages
tdl chat export -c @channel_name -T last -i 20 --all --with-content -o /tmp/tdl-export.json

# Read and process the export
cat /tmp/tdl-export.json

The export JSON contains message objects with fields like id, date, message (text content), from_id, views, and media metadata.

Step 5: Present the Results

After fetching data, present it clearly for financial research:

1. Summarize key messages — highlight the most relevant news or market updates
2. Include timestamps — show when each message was posted
3. Group by topic — if multiple channels, organize by theme (macro, earnings, crypto, etc.)
4. Flag actionable information — note breaking news, price targets, earnings surprises
5. Provide channel context — mention which channel/group each message came from
6. For channel lists, show channel name, member count, and type

Step 6: Diagnostics

If something isn't working:

Reference Files

• references/commands.md — Complete tdl command reference for reading channels and exporting messages

Read the reference file when you need exact command syntax or detailed flag documentation.