patch: Instructions file for Claude Code

CLAUDE.md

patch CLAUDE.md is an instructions file for Claude Code from recflare/patch. It costs 5,583 tokens per session, scanned A, original, MIT.

Repository instructions for patching a BepInEx 6 Harmony plugin, which changes how the Rec Room game client connects to a self-hosted server. They cover building, deploying, and avoiding known problems.

In plain words
What is it for?
Use them when modifying, compiling, or deploying this Rec Room plugin, especially after a game-version upgrade or when verifying an image signature.
Why use it?
They explain game-specific build requirements and errors that could otherwise look like code failures, such as a locked plugin file or missing generated interop files.

Instructions file for Claude Code

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

This is recflare/patch's own configuration. It tells Claude Code how to work on patch 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 patch configures →

Reuse

Borrowing it

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

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 patch CLAUDE.md

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

Your own site · 80×15
<a href="https://agentmods.dev/instructions/recflare/patch/claude-md"><img src="https://agentmods.dev/badge/instructions/recflare/patch/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 5,583 This file is loaded in full into every session.
When invoked 5,583 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 $0.05583 $0.05583
Opus 5.5 $0.02233 $0.02233
Sonnet 5.5 $0.01117 $0.01117
Haiku 4.5 $0.00558 $0.00558

Measured 2mo ago against content hash 7ec187965919, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

patch 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 · 310 lines

How it starts

The opening of the file, as written. The whole thing — 310 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

Guidance for working in this repo. This is a BepInEx 6 (IL2CPP) Harmony plugin that points the Rec Room client at a self-hosted server. Read the README for the user-facing overview; this file is the stuff you only learn by getting burned.

Build & deploy

dotnet build -c Debug -p:GamePath="C:\Games\depots\471711\23191908"
  • GamePath points at the Rec Room install root. It's normally set in the gitignored GamePath.props (see GamePath.props.example); the -p:GamePath=... override is handy for one-offs.
  • The project references ~300 interop DLLs from $(GamePath)\BepInEx\interop. Those are generated by Il2CppInterop the first time the game runs under BepInEx — if they're missing, launch the game once.
  • A post-build DeployPlugin target copies RecNetPlugin.dll into $(GamePath)\BepInEx\plugins\. The copy fails while Rec Room is running (the DLL is locked) — that's an MSB3027 error, not a compile error. Close the game and rebuild. The DLL is also always left in bin/Debug/net6.0/.

Hard-won gotchas (read before patching anything)

  1. Interop assemblies are stubs. The real code is native. The DLLs under BepInEx/interop are Il2CppInterop proxies — method bodies just marshal into GameAssembly.dll. dnSpy / managed decompilation of the interop shows no real logic. You cannot read the actual algorithms statically; you learn behavior by patching + logging at runtime.

  2. Obfuscated names differ per game build. Rec Room's type/method names are obfuscated (PGECJHKNIEN, MDBMGOBECDJ, cm_did_ppk, etc.). A dnSpy dump from some build will not necessarily match the interop you compile against. Real example from the DUID work: a dnSpy dump called the method CheckForMismatch and the pref field DBAIOPIEJNC, but in our interop the method is CheckForDUIDMismatch and neither DBAIOPIEJNC nor MDBMGOBECDJ exist at all. Always resolve members against the interop DLLs you actually build against (see the Cecil snippet below), never against a dump from an unknown build.

Read the full file on GitHub · 310 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 · 310 lines · 5,583 tokens per session scan A 7ec187965919

Subscribe to this mod's changes

patch CLAUDE.md is an instructions file published in the GitHub repository recflare/patch (7 stars, last pushed 2mo ago), licensed MIT. It adds 5,583 tokens to every session, about $0.0223 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 2026-08-31.

Related

patch CLAUDE.md compared with instructions from other repositories

patch CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
patch CLAUDE.mdrecflare/patchInstructions file72mo agoMITA5,583Claude Code
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file143ktodayMITA7,461Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA5,685GitHub Copilot
langchain AGENTS.mdlangchain-ai/langchainInstructions file149ktodayMITA4,680Codex, OpenCode
deepseek-harness AGENTS.mddeepseek-ai/deepseek-harnessInstructions file279ktodayMITA4,325Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file195ktodayMITA3,086GitHub Copilot