mcp-tg: Instructions file for Claude Code

CLAUDE.md

mcp-tg CLAUDE.md is an instructions file for Claude Code from lexfrei/mcp-tg. It costs 21,075 tokens per session, scanned B, original, BSD-3-Clause.

A development guide for mcp-tg, a server that lets compatible AI agents use Telegram's client API to access Telegram features. It also describes the Go build and test commands and the documentation workflow.

In plain words
What is it for?
Use it when developing mcp-tg, running its checks, maintaining MkDocs documentation, or reviewing its Telegram client architecture.
Why use it?
It explains how to build, test, lint, document, and publish changes in a project with many agent-accessible Telegram tools.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: mentions CLAUDE.md.

This is lexfrei/mcp-tg's own configuration. It tells Claude Code how to work on mcp-tg itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything mcp-tg configures →

Reuse

Borrowing it

Nothing to install: this file belongs to lexfrei/mcp-tg. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/lexfrei/mcp-tg/master/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/lexfrei/mcp-tg

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for mcp-tg CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/lexfrei/mcp-tg/claude-md/github.svg)](https://agentmods.dev/instructions/lexfrei/mcp-tg/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/lexfrei/mcp-tg/claude-md"><img src="https://agentmods.dev/badge/instructions/lexfrei/mcp-tg/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for mcp-tg CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/lexfrei/mcp-tg/claude-md"><img src="https://agentmods.dev/badge/instructions/lexfrei/mcp-tg/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 21,075 This file is loaded in full into every session.
When invoked 21,075 The same file — it is already loaded in full.
Security scan B 2 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.21075 $0.21075
Opus 5.5 $0.08430 $0.08430
Sonnet 5.5 $0.04215 $0.04215
Haiku 4.5 $0.02107 $0.02107

Measured 1mo ago against content hash de1f91c84948, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade B, and why

mcp-tg CLAUDE.md scanned grade B with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

GoReleaser's `nfpms` block builds a `.deb` per Linux arch out of the same builds and attaches it to the release; `lexfrei/apt` pulls it from there on its next run and serves it from <https://apt.lexfrei.dev>, so nothing

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -sS "https://raw.githubusercontent.com/DavidAnson/markdownlint-cli2-action/$(grep -oE 'markdownlint-cli2-action@[0-9a-f]{40}' .github/workflows/docs-validate.yaml | cut -d@ -f2)/package.json" | grep '"markdownlint-c
CLAUDE.md · 373 lines

How it starts

The opening of the file, as written. The whole thing — 373 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — Development Guide for mcp-tg

What is this

MCP server for Telegram Client API (MTProto, not Bot API). Uses gotd/td for protocol, exposes 80 tools + resources + prompts via MCP.

Build & Test

go build ./cmd/mcp-tg
go test -race ./...
golangci-lint run

Documentation

User-facing documentation lives in docs/, is built with MkDocs Material (mkdocs.yml) from the fully locked dependencies in requirements-docs.txt (compiled from requirements-docs.in), and follows the same shape as this owner's other docs sites: sections with their own index.md (navigation.indexes), a stylesheet, and a docs.yaml / docs-validate.yaml pair. .github/workflows/docs.yaml publishes to https://mcp-tg.lexfrei.dev on every push to master; .github/workflows/docs-validate.yaml runs markdownlint-cli2 and the same mkdocs build --strict on every PR (TestDocsGates_RunOnPullRequests). README.md is a landing page only — install, quickstart, links — and anything longer belongs on a docs page.

That PR check must be in master's required status checks, and is — today they are Lint, Test, Build (amd64), Build (arm64), Validate Documentation and Release Config. Keep it there: nothing in the workflow graph depends on Validate Documentation, so were it dropped, a PR that breaks the site would merge green, docs.yaml would fail on master, and the site would quietly stop updating with the last good build still served. The list lives in the repository's branch-protection settings, outside version control like the Pages config below, so it can drift from this paragraph with nothing red; gh api repos/lexfrei/mcp-tg/branches/master/protection --jq '.required_status_checks.contexts' is the source of truth. Relative links and anchors inside docs/ ([Peers](peers.md), ../building.md#transport-modes) are checked by that build and by nothing else — TestDocsSiteURLs_ResolveToPages reads absolute URLs only. This is why docs-validate.yaml carries no paths: filter, unlike the reference it is modelled on: GitHub leaves a skipped workflow's check Pending forever and blocks a merge on a required check that never reports, so a filtered workflow can never be required at all.

Read the full file on GitHub · 373 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 1mo ago Changed · +2 lines · +268 tokens per session scan A → B de1f91c84948
  2. 1mo ago Changed · +1 lines · +177 tokens per session 8679a9b35df6
  3. 1mo ago Changed · +15 lines · +1,127 tokens per session 34c6c2dcd331
  4. 1mo ago Changed · +3 lines · +1,521 tokens per session e6477d35b5aa
  5. 2mo ago First seen · 352 lines · 17,982 tokens per session scan A 0b3e27e20af7

Subscribe to this mod's changes

mcp-tg CLAUDE.md is an instructions file published in the GitHub repository lexfrei/mcp-tg (9 stars, last pushed yesterday), licensed BSD-3-Clause. It adds 21,075 tokens to every session, about $0.0843 per session on Opus 5.5. A static security scan graded it B with 2 findings (asks for root, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

mcp-tg CLAUDE.md compared with instructions from other repositories

mcp-tg CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
mcp-tg CLAUDE.mdlexfrei/mcp-tgInstructions file9yesterdayBSD-3-ClauseB21,075Claude Code
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143kyesterdayMITA7,461Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchainInstructions file149kyesterdayMITA4,680Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA5,516GitHub Copilot
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195kyesterdayMITA3,093GitHub Copilot
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harnessInstructions file274k2d agoMITA4,313Codex, OpenCode