What is nvim-ui-mcp CLAUDE.md?

A set of instructions for an MCP server that lets an AI agent view and control Neovim through its rendered screen. Neovim is a text editor; the server can account for visible elements such as floating windows, completion menus, highlights, the cursor, and the current mode.

Use it when building or testing an agent that interacts with Neovim's user interface, including its screen rendering and layered TypeScript implementation.

It helps an agent check what is actually visible in the editor instead of relying only on internal editor state. The intended workflow is to observe the screen, act, observe again, and verify the result.

How kjoonha/nvim-ui-mcp configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to kjoonha/nvim-ui-mcp. 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/kjoonha/nvim-ui-mcp/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/kjoonha/nvim-ui-mcp

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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/kjoonha/nvim-ui-mcp/claude-md.svg)](https://agentmods.dev/instructions/kjoonha/nvim-ui-mcp/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/kjoonha/nvim-ui-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/kjoonha/nvim-ui-mcp/claude-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 1,487 This file is loaded in full into every session.

When invoked 1,487 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.

Model	Per session	Once invoked
Fable 5.1	$0.01487	$0.01487
Opus 5	$0.00744	$0.00744
Sonnet 5	$0.00297	$0.00297
Haiku 4.5	$0.00149	$0.00149

Measured 6d ago against content hash ab21ceb916e4, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

nvim-ui-mcp 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 6d 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 · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

nvim-ui-mcp

Playwright for Neovim agents. MCP server exposing the rendered Neovim UI (not just buffer/API state) so AI agents can observe→act→observe→verify against real screen output — floats, completion menus, highlights, cursor, mode.

Full architecture/rationale: DESIGN.md (read it before touching RPC/screen/session layers). Original requirements: .omc/specs/deep-interview-nvim-ui-mcp.md. Consensus plan: .omc/plans/nvim-ui-mcp-plan.md.

Three principles (drive every design decision here)

Rendered UI is source of truth.
Structured Neovim state augments the rendered UI, never replaces it.
Primary loop: observe → act → observe → verify.

Stack

TypeScript/Node 22+, @modelcontextprotocol/server (stdio transport), neovim npm client (msgpack-RPC, nvim_ui_attach + ext_linegrid), Zod schemas, Vitest, tsup, ESLint+typescript-eslint strict.

Architecture (layered, each independently testable)

src/rpc/        connect.ts          launch/attach to Neovim over msgpack-RPC (socket or embedded stdio)
src/ui/         events.ts, event-processor.ts    parse `redraw` batches into typed UiEvent[]
src/screen/     cell.ts, shadow-screen.ts, serializer.ts, diff.ts   pure grid model, no transport/MCP deps
src/session/    session.ts, manager.ts, observation.ts, wait.ts    ties RPC+screen together, session lifecycle
src/mcp/        server.ts, stdio.ts, tools/*.ts    MCP tool registration + handlers

Data flow: rpc/connect → RPC client emits redraw notifications → ui/event-processor parses into UiEvent[] → screen/shadow-screen mutates grid state → screen/serializer renders text → mcp/tools/* package it for the agent.

ShadowScreen consistency gate

Screen is only safe to read when consistent === true. Mutation sets consistent = false and clears dirty rows; the flush event (last event in a redraw batch) sets consistent = true. Never read mid-batch.

Session lifecycle

SessionManager (unbounded Map, no cap/idle-timeout by design choice — flagged nice-to-have, not fixed) owns NeovimSession instances. Two modes:

test mode (nvim_launch): server spawns+owns nvim --embed, nvim_close may kill it.
attach mode (nvim_attach): connects to an existing --listen socket/named-pipe; nvim_close refuses these (not ours to kill).

Read the full file on GitHub · 68 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.

6d ago First seen · 68 lines · 1,487 tokens per session scan A ab21ceb916e4

Subscribe to this mod's changes

nvim-ui-mcp CLAUDE.md is an instructions file published in the GitHub repository kjoonha/nvim-ui-mcp (0 stars, last pushed 26d ago), licensed MIT. It adds 1,487 tokens to every session, about $0.0074 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 2026-08-31.

Frontend

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

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

nvim-ui-mcp: Instructions file for Claude Code