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/u9401066/rootcause-mcp/copilot-instructionsgit clone --depth 1 https://github.com/u9401066/rootcause-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/u9401066/rootcause-mcp/copilot-instructions)<a href="https://agentmods.dev/instructions/u9401066/rootcause-mcp/copilot-instructions"><img src="https://agentmods.dev/badge/instructions/u9401066/rootcause-mcp/copilot-instructions.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 | $0.02974 | $0.02974 |
| Opus 5 | $0.01487 | $0.01487 |
| Sonnet 5 | $0.00595 | $0.00595 |
| Haiku 4.5 | $0.00297 | $0.00297 |
Grade A, and why
rootcause-mcp copilot-instructions.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.
How it starts
The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Copilot 自定義指令
此文件為 VS Code GitHub Copilot 及 Claude Code 提供專案上下文與操作規範。
專案概述
這是 RootCause MCP 2.x 臨床推理與醫療根因分析專案。它以 MCP SDK 2 提供 46 個 discrete tools 或 8 個 condensed facade tools,讓不同 host agent 將多來源、去識別化病歷整理成具來源血緣的時間線、鑑別診斷、evidence links、認知偏誤審查、Fishbone/Why/HFACS 保守因果稽核與標準化交接產物。
RootCause 是回溯性決策支援與稽核 harness,不是自主診斷或治療系統。所有 case workflow 必須遵守 exact provenance、PHI 最小化、保守因果判定及 qualified-human review。
MCP 本身不會思考或下診斷;host Agent 負責臨床推論,RootCause MCP 只保存、驗證與約束可稽核產物。對醫師的敘述使用繁體中文,diagnosis、test、drug、procedure 等標準名稱保留 English;首次出現的既定縮寫可附繁中對照,來源中的原文與單位不得翻譯改寫。
建立 DDx 時,先以 phenotype、time course 與不同 mechanism category 展開最大合理廣度,再選 syndrome-appropriate framework 逐一審查所有 required cells。Final PRIMARY breadth audit 不得留下 NOT_ASSESSED;REVIEWED_INSUFFICIENT_DATA 必須保留 unknowns 與 typed planned discriminators,不能視為排除。三個不重複候選是 deterministic finalization floor,不是推理目標或上限;同時避免同義重複與無限 laundry list。每個 active candidate 都要說明 why considered、source-linked support/refute/neutral evidence、unknown、discriminating test 與 qualitative certainty,並用具理由的 mutation 明確選擇 leading diagnosis。unknown 不是 negative finding;不得捏造 probability 或 LR,LR=1.0 只能表示 neutral/quantitatively unknown。非中性 LR 必須另外連到 verified LITERATURE calibration evidence,不能只寫一個看似合法的 citation 字串。
時間必須保留來源精度:只有來源自帶 offset 的 instant 可排序或支持 temporality;date、range、relative、unknown 都是合法 final 狀態,但不可被補成假時間。Pinned manifest 的 identity/digest 不可改;抽取後以 allowlisted reviewer 的 rc_adjudicate_source append review/independence event。每個 Fishbone cause 也必須由 allowlisted reviewer persist HFACS CONFIRMED 或 NOT_APPLICABLE disposition,不能把 suggestion 或 rc_add_cause 的任意 code 當人工確認。
開發哲學 💡
「想要寫文件的時候,就更新 Memory Bank 吧!」
「想要零散測試的時候,就寫測試檔案進 tests/ 資料夾吧!」
- 不要另開檔案寫筆記,直接寫進 Memory Bank
- 今天的零散測試,就是明天的回歸測試
法規層級
CONSTITUTION.md ← 最高原則(不可違反)
│
├── .github/bylaws/ ← 子法(細則規範)
│ ├── ddd-architecture.md
│ ├── git-workflow.md
│ ├── python-environment.md
│ └── memory-bank.md
│
└── .codex/skills/rootcause-clinical-reasoning-harness/ ← RootCause case workflow canonical skill
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 · 238 lines · 2,974 tokens per session scan A b9ee079be6d5
rootcause-mcp copilot-instructions.md is an instructions file published in the GitHub repository u9401066/rootcause-mcp (0 stars, last pushed 15d ago), licensed Apache-2.0. It adds 2,974 tokens to every session, about $0.0149 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
medical-calc-mcp copilot-instructions.md
Instructions for u9401066/medical-calc-mcp, covering copilot instructions for medical calculator mcp, 📊 專案概況, 🛠️ 開發環境, 安裝依賴 and 執行測試.
sagemath-mcp CLAUDE.md
Instructions for XBP-Europe/sagemath-mcp, covering claude.md, project, commands, linting and testing.
blender-mcp copilot-instructions.md
Copilot instructions for sandraschi/blender-mcp: You have access to a Blender 3D automation server. You can control Blender scenes, objects, materials, rendering, and physics through MCP tools.
ai4science-studio CLAUDE.md
Instructions for AMDResearch/ai4science-studio, covering claude.md, repository purpose, agent entry points (read these first), directory layout and model slug rule.
leanforge-mcp CLAUDE.md
Claude Code instructions for sandraschi/leanforge-mcp, covering claude.md -- leanforge-mcp, what this repo does, key concepts, when working on agent.py and when working on leanclient.py.
orphanet-link CLAUDE.md
Claude Code instructions for berntpopp/orphanet-link, covering claude.md, essentials and common commands.