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.
npx agentmods add instructions/fc0web/lab-notebook-mcp/claude-mdgit clone --depth 1 https://github.com/fc0web/lab-notebook-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/fc0web/lab-notebook-mcp/claude-md)<a href="https://agentmods.dev/instructions/fc0web/lab-notebook-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/fc0web/lab-notebook-mcp/claude-md.svg" alt="Measured on agentmods" 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 | $0.00808 | $0.00808 |
| Opus 5 | $0.00404 | $0.00404 |
| Sonnet 5 | $0.00162 | $0.00162 |
| Haiku 4.5 | $0.00081 | $0.00081 |
Grade A, and why
lab-notebook-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 3d 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.
What it actually says
lab-notebook-mcp
測定の一次資料を永続化する MCP サーバー。benchtop(測定)と rei-aios(理論)を つなぐ共有レイヤー。設計の詳細は docs/design.md を参照。
不変の原則(変更しないこと)
- Append-only: 削除ツールを追加しない。訂正は supersede で表現する。
- モードゲーティング: ro モードでは書き込みツールを登録しない。 「呼ばれてから拒否」ではなく「存在させない」。
- 任意 SQL ツールを作らない: 必要な問い合わせは型付き専用ツールで追加する。
- 外部由来テキストは必ず でラップする。 provenance='external' のレコードを裸で返さない。
- 測定器に触らない: benchtop の責務。ここから機器へコマンドを送らない。
- アーティファクトの中身を返さない: メタデータとパスのみ。
運用ルール
run.operatorにはイニシャルかハンドル名のみ。氏名を書かない。note.bodyに顧客名・契約情報・個人情報を書かない。- 単位は experiment 内で統一する(mV と V を混在させない)。
開発
# 依存インストール (uv 使用時)
uv sync
uv run pytest
# または pip
python -m venv .venv
.venv/Scripts/activate # Windows
# source .venv/bin/activate # Linux/Mac
pip install -e ".[dev]"
pytest
# サーバー起動 (開発時、 rw モード)
LAB_NOTEBOOK_MODE=rw LAB_NOTEBOOK_ROOT=./_dev python -m lab_notebook
役割分担
- 本サーバー: 事実の永続化・検索・比較
- benchtop: 測定器の制御と生データ取得
- rei-aios: 仮説生成と理論構築(本サーバーへは ro でのみ接続)
Cross-pollination with Rei stack
本 project は Rei stack の 4 systems (rei-aios / benchtop / mcp-lens / rei-automator-mcp) を 5 systems 化する 拡張だが、 chat-Claude 前 turn 「+4 MCP → AI 選択劣化」 warning に 抵触しない設計:
- 責務が clean separation (benchtop=I/O, lab-notebook=永続化, rei-aios=推論)
- 各 MCP が 独立して valid (相互 HTTP/IPC 依存なし、 ID sharing のみ)
- ro モードで rei-aios 自律ループから 参照 = write 権限 は Claude Code / Cowork Claude のみ
関連 arc:
- Rei-Automator STEP 1340 audit-log.ts (2026-08-17、 hash chain primitive origin)
- benchtop v0.3.0 audit chain port (2026-08-17)
- benchtop v0.4.0 experiment notebook 統合 (2026-08-17、 本 project と 責務 overlap あり、 「quick session-level notes」 vs 「structured multi-session records」 使い分け)
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.
- 3d ago First seen · 60 lines · 808 tokens per session scan A f9ad38298a14
lab-notebook-mcp CLAUDE.md is an instructions file published in the GitHub repository fc0web/lab-notebook-mcp (0 stars, last pushed 12d ago), licensed MIT. It adds 808 tokens to every session, about $0.0040 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.
Other instructions, from other repositories
garmin-health-data CLAUDE.md
Instructions for diegoscarabelli/garmin-health-data, covering claude.md, readme.md, database schema architecture, schema documentation and formatting ddl files.
Biomedical-Agent-DB-Karpathy CLAUDE.md
Instructions for stanveer/Biomedical-Agent-DB-Karpathy, covering claude.md, 1. normalize, don't guess, 2. keep entity types separate, 3. provenance first and 4. minimal, verifiable schema changes.
mcp-openfoodtox AGENTS.md
Instructions for spyrosze/mcp-openfoodtox, covering efsa openfoodtox mcp server - project instructions, project overview, dataset structure, core tables (sqlite) and key relationships.
chembl-mcp-server AGENTS.md
AGENTS.md instructions for cyanheads/chembl-mcp-server, covering developer protocol, server surface, core rules, patterns and tool.
odsbox-jaquel-mcp AGENTS.md
AGENTS.md instructions for totonga/odsbox-jaquel-mcp: This repository contains the code of an FastMCP agent that can be used to query the odsboxpython package to access ASAM ODS servers. Its setup to use modern Python tools.
braina GEMINI.md
Instructions for brainets/braina, covering project: braina (brain interaction analysis), 1. project context & purpose, 2. commands, verify environment (all core dependencies) and run the verification test suite for frites + hoi.