Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/limehq/munkel/claude-mdgit clone --depth 1 https://github.com/limehq/munkelWrote 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.
[](https://agentmods.dev/instructions/limehq/munkel/claude-md)<a href="https://agentmods.dev/instructions/limehq/munkel/claude-md"><img src="https://agentmods.dev/badge/instructions/limehq/munkel/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.02131 | $0.02131 |
| Opus 5 | $0.01066 | $0.01066 |
| Sonnet 5 | $0.00426 | $0.00426 |
| Haiku 4.5 | $0.00213 | $0.00213 |
Grade A, and why
munkel CLAUDE.md scanned grade A with 0 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 4d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Overview
Munkel: ephemeral, end-to-end-encrypted messages between friends ("channels") that slide out of the MacBook notch. No accounts, no message storage.
docs/ARCHITECTURE.md is the authoritative, source-linked map of every component
and the data flow — read it before non-trivial work instead of re-deriving the
big picture here.
Monorepo layout
Bun workspaces + Turborepo. Four apps under apps/:
apps/macos— Swift/SwiftUI menu-bar app. Owns all crypto and relay connections; the CLI and the UI are clients of it. Two SwiftPM targets:MunkelKit(reusable lib: crypto, protocol, relay/blob clients, GitHub auth) andMunkelApp(menu-bar UI, notch presentation, control socket).apps/cli— themunkelCLI (Bun/TS). Thin and send-only: no crypto, no relay connection; serializes one JSON request over the app's Unix socket.apps/server— Cloudflare Worker relay (munkel-relay, Hono router) + one Durable Object per channel (partyserver, WebSocket hibernation, no storage).apps/landing— TanStack Start (React + Tailwind v4) on a Worker; marketing only, no access to relay traffic or keys.
Commands
Everything runs from the repo root (Turbo fans out across packages):
bun install
bun run test # Swift MunkelKit tests + CLI + server unit/e2e
bun run build
bun run typecheck
bun run dev deliberately excludes @munkel/macos (it needs the Swift
toolchain) — run the macOS app with its own command below.
Per-package dev servers (--filter):
- Relay:
bunx turbo dev --filter=@munkel/server→ws://127.0.0.1:8787 - Landing:
bunx turbo dev --filter=@munkel/landing→http://localhost:3000 - macOS app:
cd apps/macos && bun run dev(builds + runs the Munkel Dev variant)
Single tests:
- Server (vitest):
cd apps/server && bunx vitest run test/blob.test.ts(or-t "<name>"); e2e only:bun run test:e2e - CLI (bun):
cd apps/cli && bun test test/munkel.test.ts - Swift:
cd apps/macos && swift test --filter CryptoTests
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.
- 4d ago First seen · 150 lines · 2,131 tokens per session scan A 6b9beb157ac3
munkel CLAUDE.md is an instructions file published in the GitHub repository limehq/munkel (22 stars, last pushed 4d ago), licensed MIT. It adds 2,131 tokens to every session, about $0.0107 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
stoop CLAUDE.md
Instructions for mislavjc/stoop, covering stoop agent notes, workspace layout, commands, effect conventions and schema contracts.
stoop AGENTS.md
Instructions for mislavjc/stoop, covering release workflow, write changelog files, example, fix button hover state and package references.
vibesdk AGENTS.md
AGENTS.md instructions for cloudflare/vibesdk, covering agents.md, tooling, verification, frontend ui and frontend data fetching.
vibesdk CLAUDE.md
Claude Code instructions for cloudflare/vibesdk, covering claude.md, communication style, project overview, key architectural patterns and common development tasks.
OnlySwitch AGENTS.md
Instructions for jacklandrin/OnlySwitch, covering onlyswitch agent guidance, workflow, context routing and project-specific working rules.
Applite CLAUDE.md
Instructions for milanvarady/Applite, covering claude.md, project overview, build & run, architecture and data flow.