RenderDocMCP CLAUDE.md

Project instructions for RenderDocMCP, a server that lets an AI assistant inspect RenderDoc graphics captures. RenderDoc is a tool for debugging rendered graphics such as DirectX 11 and 12 frames.

In plain words
What is it for?
Use it to list and open capture files, check their loading status, and retrieve filtered drawing-call information while debugging graphics.
Why use it?
It documents how the server, RenderDoc extension, and file-based communication fit together, so agents can use the project consistently.

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/halby24/renderdocmcp/claude-md
Clone the repo
git clone --depth 1 https://github.com/halby24/RenderDocMCP
Per session 1,504 This file is loaded in full into every session.
When invoked 1,504 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 $0.01504 $0.01504
Opus 5 $0.00752 $0.00752
Sonnet 5 $0.00301 $0.00301
Haiku 4.5 $0.00150 $0.00150

Measured 2d ago against content hash 48c263d9d10a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

RenderDocMCP 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 2d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

CLAUDE.md · 136 lines

How it starts

The opening of the file, as written. The whole thing — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.

RenderDoc MCP Server

RenderDoc UI拡張機能として動作するMCPサーバー。AIアシスタントがRenderDocのキャプチャデータにアクセスし、DirectX 11/12のグラフィックスデバッグを支援する。

アーキテクチャ

ハイブリッドプロセス分離方式:

Claude/AI Client (stdio)
        │
        ▼
MCP Server Process (標準Python + FastMCP 2.0)
        │ File-based IPC (%TEMP%/renderdoc_mcp/)
        ▼
RenderDoc Process (Extension + File Polling)

プロジェクト構成

RenderDocMCP/
├── mcp_server/                        # MCPサーバー
│   ├── server.py                      # FastMCPエントリーポイント
│   ├── config.py                      # 設定
│   └── bridge/
│       └── client.py                  # ファイルベースIPCクライアント
│
├── renderdoc_extension/               # RenderDoc拡張機能
│   ├── __init__.py                    # register()/unregister()
│   ├── extension.json                 # マニフェスト
│   ├── socket_server.py               # ファイルベースIPCサーバー
│   ├── request_handler.py             # リクエスト処理
│   └── renderdoc_facade.py            # RenderDoc APIラッパー
│
└── scripts/
    └── install_extension.py           # 拡張機能インストール

MCPツール

ツール名 説明
list_captures 指定ディレクトリ内の.rdcファイル一覧を取得
open_capture キャプチャファイルを開く(既存キャプチャは自動で閉じる)
get_capture_status キャプチャ読込状態確認
get_draw_calls ドローコール一覧(階層構造、フィルタリング対応)
get_frame_summary フレーム全体の統計情報(ドローコール数、マーカー一覧等)
find_draws_by_shader シェーダー名でドローコールを逆引き検索
find_draws_by_texture テクスチャ名でドローコールを逆引き検索
find_draws_by_resource リソースIDでドローコールを逆引き検索
get_draw_call_details 特定ドローコールの詳細
get_action_timings アクションのGPU実行時間を取得
get_shader_info シェーダーソース/定数バッファ
get_buffer_contents バッファデータ取得(オフセット/長さ指定可)
get_texture_info テクスチャメタデータ
get_texture_data テクスチャピクセルデータ取得(mip/slice/3Dスライス対応)
get_pipeline_state パイプライン状態全体

get_draw_calls フィルタリングオプション

get_draw_calls(
    include_children=True,      # 子アクションを含める
    marker_filter="Camera.Render",  # このマーカー配下のみ取得
    exclude_markers=["GUI.Repaint", "UIR.DrawChain"],  # 除外するマーカー
    event_id_min=7372,          # event_id範囲の開始
    event_id_max=7600,          # event_id範囲の終了
    only_actions=True,          # マーカーを除外(ドローコールのみ)
    flags_filter=["Drawcall", "Dispatch"],  # 特定フラグのみ
)

Read the full file on GitHub · 136 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. 2d ago First seen · 136 lines · 1,504 tokens per session scan A 48c263d9d10a

Subscribe to this mod's changes

RenderDocMCP CLAUDE.md is an instructions file published in the GitHub repository halby24/RenderDocMCP (486 stars, last pushed 7mo ago), licensed MIT. It adds 1,504 tokens to every session, about $0.0075 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.