mitos: Instructions file for Claude Code

CLAUDE.md

mitos CLAUDE.md is an instructions file for Claude Code from mitos-run/mitos. It costs 2,367 tokens per session, scanned A, original, Apache-2.0.

Repository instructions for Mitos Run, a system that creates and manages isolated computing sandboxes using Kubernetes and lightweight virtual machines.

In plain words
What is it for?
Use them when working on sandbox controllers, virtual-machine forking, guest agents, REST or gRPC APIs, the Python SDK, benchmarks, or threat-model updates.
Why use it?
They explain the project’s architecture, priorities, security rules, API target, and verification requirements before code is changed.

Instructions file for Claude Code

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

This is mitos-run/mitos's own configuration. It tells Claude Code how to work on mitos 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 mitos configures →

Reuse

Borrowing it

Nothing to install: this file belongs to mitos-run/mitos. 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/mitos-run/mitos/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/mitos-run/mitos

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 mitos CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/mitos-run/mitos/claude-md/github.svg)](https://agentmods.dev/instructions/mitos-run/mitos/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/mitos-run/mitos/claude-md"><img src="https://agentmods.dev/badge/instructions/mitos-run/mitos/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 mitos CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/mitos-run/mitos/claude-md"><img src="https://agentmods.dev/badge/instructions/mitos-run/mitos/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 2,367 This file is loaded in full into every session.
When invoked 2,367 The same file — it is already loaded in full.
Security scan A 0 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 $0.02367 $0.02367
Opus 5.5 $0.00947 $0.00947
Sonnet 5.5 $0.00473 $0.00473
Haiku 4.5 $0.00237 $0.00237

Measured 2mo ago against content hash 23d552c185dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

mitos 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 2mo 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.

CLAUDE.md · 139 lines

How it starts

The opening of the file, as written. The whole thing — 139 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

Project Overview

Snapshot-fork sandboxes for AI agents on Kubernetes. The system boots Firecracker microVMs, forks them via copy-on-write snapshots, and exposes the whole lifecycle through declarative CRDs (SandboxPool, Sandbox, Workspace) in API group mitos.run/v1.

Components:

  • controller (Deployment): reconciles the CRDs, selects nodes, drives forkd.
  • forkd (DaemonSet): per-node fork daemon; gRPC on :9090 for the controller, HTTP sandbox API on :9091 for exec and file traffic.
  • guest agent (PID 1 in the VM): speaks the vsock protocol for exec, files, env, and fork notifications.
  • sandbox-server (standalone): the same engine behind a plain REST API, no Kubernetes required.
  • Python SDK (sdk/python): client for both k8s mode and sandbox-server mode.

ROADMAP.md is the priority order for all work. docs/api/v2-spec.md is the target API.

Operating Principles

These outrank convenience:

  1. No unverified claims. Every public number must be reproducible from bench/ or it does not get written.
  2. Security findings block features. The threat model (docs/threat-model.md) must be updated in the same PR whenever the security surface moves.
  3. Honest Kubernetes semantics. Sandboxes are not pods; never imply pod-scoped mechanisms (NetworkPolicy, ResourceQuota, PSA) govern them.
  4. Boring failure behavior. Every component defines what happens on crash, node loss, slow etcd, and capacity exhaustion.
  5. Bare metal is a first-class target.
  6. Experience is DNA. Every user-facing surface follows the journey rules: no dead ends, simple surface with depth one click down, intent-shaped aha. See docs/superpowers/specs/2026-06-27-hosted-launch-journey-design.md.
  7. Self-host is first-class (Apache-2.0). mitos is open source and people run it themselves; the self-hosted experience is a peer of the hosted one, never an afterthought to the SaaS. One console image and SPA serve both, differing ONLY by the capabilities document (GET /console/capabilities); never fork an edition build. Every hosted-only surface (self-serve signup, billing, credits and top-up, the allowlist gate, Paddle, abuse email checks) MUST be capability-gated so a self-hoster gets a clean, complete product: no dead links, no empty billing panels, no "add credits" or signup prompts. The journey rules apply to the community first-run too; it points at the SDK/CLI and a real first success, never at a paywall.

Read the full file on GitHub · 139 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. 2mo ago First seen · 139 lines · 2,367 tokens per session scan A 23d552c185dd

Subscribe to this mod's changes

mitos CLAUDE.md is an instructions file published in the GitHub repository mitos-run/mitos (91 stars, last pushed 3mo ago), licensed Apache-2.0. It adds 2,367 tokens to every session, about $0.0095 per session on Opus 5.5. 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.

Related

mitos CLAUDE.md compared with instructions from other repositories

mitos CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
mitos CLAUDE.mdmitos-run/mitosInstructions file913mo agoApache-2.0A2,367Claude 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