wingthing CLAUDE.md

wingthing CLAUDE.md is an instructions file for coding agents from ehrlich-b/wingthing. It costs 6,225 tokens per session, scanned A, original, MIT.

A set of instructions for Wingthing, a local control system for running and supervising AI agents through a command line, browser, or MCP. MCP is a standard way for language models to call tools.

In plain words
What is it for?
Use it to start Claude Code sessions in sandboxes, let a local language model supervise agents, connect a computer to the Wingthing portal, or run its self-hosted gateway.
Why use it?
It explains how the system organizes agents, sessions, permissions, and records of actions in one place. The input does not describe a separate coding feature.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/ehrlich-b/wingthing/claude-md
Clone the repo
git clone --depth 1 https://github.com/ehrlich-b/wingthing

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 wingthing CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/ehrlich-b/wingthing/claude-md.svg)](https://agentmods.dev/instructions/ehrlich-b/wingthing/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/ehrlich-b/wingthing/claude-md"><img src="https://agentmods.dev/badge/instructions/ehrlich-b/wingthing/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 6,225 This file is loaded in full into every session.
When invoked 6,225 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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 $0.06225 $0.06225
Opus 5 $0.03112 $0.03112
Sonnet 5 $0.01245 $0.01245
Haiku 4.5 $0.00622 $0.00622

Measured 4d ago against content hash 233db9e23f41, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

wingthing 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.

CLAUDE.md · 424 lines

How it starts

The opening of the file, as written. The whole thing — 424 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Wingthing

What This Is

wt puts AI agent runtimes behind one local control plane. A person uses that runtime through the CLI or browser. An LLM uses it through MCP. The primary use case is all of a person's agents in one inventory, for that person and their agents to operate together. Skills are a secondary feature.

  • wt egg claude -- run Claude Code in a per-session sandbox with PTY persistence
  • wt mcp stdio --client codex -- let a local LLM start and supervise agents
  • wt start -- connect your machine to the relay, access from app.wingthing.ai
  • wt roost start -- self-hosted portal/gateway plus an embedded wing
  • wt serve -- gateway only (web UI, WebSocket relay, skill registry), HTTP + SQLite

Current Push: One Portal for People and LLMs

An AI must be able to orchestrate wingthing as easily as a human can. This is the primary focus of the current work. Anything a human can do from the terminal or the browser, a model must be able to do through a typed interface with the same authority model and the same audit trail.

The rule: if you ship a capability only a human can drive, it is unfinished.

"One portal" means one resource inventory, authority model, and lifecycle contract. It does not mean every process or file moves to one machine.

  • A portal is the client-facing inventory and control surface.
  • A wing is the execution runtime and source of truth for local state.
  • A session is a persistent interactive terminal.
  • A run is a supervised headless task with semantic state.
  • An egg is the per-session process, PTY, and sandbox boundary.
  • A roost is the self-hosted bundle from wt roost start: portal/gateway plus an embedded wing.

Browser and MCP adapters must address the same qualified resources. Cross-wing resources need stable portal, wing, kind, and object IDs. Do not add mutable "current wing" state to a server-side MCP session.

Every remote workflow must place execution, workspace, display, credentials, and durable memory explicitly. Current cwd values refer to existing paths on the selected wing. Wingthing does not yet synchronize code or memory between wings, and docs must not imply that it does.

Read the full file on GitHub · 424 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. 4d ago First seen · 424 lines · 6,225 tokens per session scan A 233db9e23f41

Subscribe to this mod's changes

wingthing CLAUDE.md is an instructions file published in the GitHub repository ehrlich-b/wingthing (53 stars, last pushed 5d ago), licensed MIT. It adds 6,225 tokens to every session, about $0.0311 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.

Related

Other instructions, from other repositories

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

spec-kit AGENTS.md

AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,104 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,345 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).

microsoft/vscode · 5,001 tokens

next.js AGENTS.md

Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens