Borrowing it
Nothing to install: this file belongs to Automaat/lightroom-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.
curl -O https://raw.githubusercontent.com/Automaat/lightroom-mcp/main/CLAUDE.mdgit clone --depth 1 https://github.com/Automaat/lightroom-mcpWrote 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/automaat/lightroom-mcp/claude-md)<a href="https://agentmods.dev/instructions/automaat/lightroom-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/automaat/lightroom-mcp/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.
<a href="https://agentmods.dev/instructions/automaat/lightroom-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/automaat/lightroom-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.04585 | $0.04585 |
| Opus 5.5 | $0.01834 | $0.01834 |
| Sonnet 5.5 | $0.00917 | $0.00917 |
| Haiku 4.5 | $0.00458 | $0.00458 |
Grade A, and why
lightroom-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 today.
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 — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Lightroom MCP
MCP server bridging Claude to Adobe Lightroom Classic.
Layout
server/— TypeScript MCP server (ESM, NodeNext). Stdio transport for Claude + raw TCP client to plugin sockets.plugin/LightroomMCP.lrplugin/— Lua plugin loaded by Lightroom Classic.manual-test.mjs— direct TCP probe (bypasses MCP) for validating plugin dispatch.
Architecture (read before changing transport)
Plugin opens two LrSocket binds as servers; MCP server connects to both.
- Plugin :58763 (default) in
mode='receive'— server writes line-delimited JSON requests - Plugin :58764 (default) in
mode='send'— server reads line-delimited JSON responses - Frame:
\nterminator on every message (LrSocket buffers until newline) - Plugin allows one client per port at a time. MCP server holds a persistent connection.
- Hosts spawn several bridges at once (Claude Desktop: chat + Cowork/Code shared pool, #238).
Bridge(server/src/bridge.ts) elects one per port pair via the instance lock: the leader owns the plugin sockets and serves tool calls over local IPC (broker.ts: Unix socket in~/.config/lightroom-mcp/, named pipe on Windows, plugin token checked); followers forward, and re-elect when the leader exits. Neverprocess.exiton lock contention. LrSocket.bindinmode='receive'has a 10s no-client timeout that firesonError. Plugin auto-calls:reconnect()from a monitor loop in response. Reconnect storms are prevented by setting flags in callbacks and acting on them in the loop (never:reconnect()synchronously fromonError).onMessageruns in non-yielding context — handler dispatch must be wrapped inLrTasks.startAsyncTasksocatalog:withReadAccessDocan yield.- Catalog queries that yield (
getTargetPhotos,findPhotos) must run outsidewithReadAccessDo. Nesting a yielding query inside the read gate deadlocks on Windows: the task yields to the UI thread holding the gate, never returns, never releases it, wedging the bridge until the server timeout (#124/#134). Only non-yielding per-photo metadata reads belong inside the gate.getAllPhotosis a non-yielding enumeration, safe either side.PhotoLookupresolves paths withfindPhotoByPath, a catalog query, so handlers resolve photo ids before opening either gate. Specs guard this viagetQueriedInsideReadAccess()inspec_helper. - Every SDK getter call (
getRawMetadata,getFormattedMetadata,LrKeywordgetters) costs ~0.1-1 ms; per-call overhead dominates handler time on large catalogs. Read photo fields for a page withcatalog:batchGetRawMetadata/batchGetFormattedMetadata(PhotoSummary), read keyword name and synonyms with onegetAttributes, and avoidLrTasks.pcallper item (~1 ms each). catalog:getPhotoByLocalIdreturns a dangling photo object, not nil, for an unknown id; its metadata reads return nil.PhotoLookupkeeps only photos whose batcheduuidread is non-nil.
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.
- today Changed · +7 lines · +488 tokens per session 63b3ecc06e07
- yesterday Changed · +4 lines · +610 tokens per session 0af0bbafbd36
- 2d ago Changed · +7 lines · +1,311 tokens per session f4b25f6930e2
- 8d ago Changed · +4 lines · +217 tokens per session ad7fb2d1db62
- 15d ago Changed · +10 lines · +434 tokens per session f4e42df875ad
- 24d ago Changed · +1 lines · +103 tokens per session f15999f56dcb
- 28d ago Changed · +54 tokens per session 3e5a10eeca69
- 1mo ago First seen · 77 lines · 1,368 tokens per session scan A 1b78980a6ef0
lightroom-mcp CLAUDE.md is an instructions file published in the GitHub repository Automaat/lightroom-mcp (158 stars, last pushed today), licensed MIT. It adds 4,585 tokens to every session, about $0.0183 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.
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).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
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.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.