I recently did some tests on two MCPs with identical functionalities. Turns out one of them has really bad performance. So I wanna share those bad MCP design patterns that cause this.

It all started when I wrote an MCP Server (MCP-A) for a to-do list app. Later, the app officially released its own MCP Server (MCP-B). Both MCPs have the same functionalities and hit the same backend API.

The experiment is set up as follows:

- Both MCP Servers connect to the same ToDo list account, and it will be reset after each test. - 40 test prompts to simulate typical use cases for these MCPs. - The test was conducted with the same model, system prompt, and Agent framework

Here are the results:

| Metric | MCP-A | MCP-B | Gap | | ------------------- | ----------- | ----------- | ----- | | Tool Desc Length | 11,464 | 3,682 | — | | Pass Rate | 36/40 (90%) | 36/40 (90%) | Same | | Total input tokens | 637,244 | 3,174,329 | 4.98× | | Total output tokens | 17,301 | 23,238 | 1.34× | | Total Agent steps | 122 | 157 | 1.29× | | Total time | 597s | 676s | 1.13× |

---

The result shows that MCP-B took 35 more ReAct loops to complete 40 test cases compared to MCP-A, which means 30% more output tokens. I examined the log and found that the root cause is poor query tool design.

Take the `search tool` for example, its job is to find a todo item in the ToDo list. In MCP-B, this tool returns this:

{ "id": "6a1916b48f08cb3a4c857ed0", "title": "buy some groceries", "url": "https://todo.example.com/tasks/6a1916b48f08cb3a4c857ed0" }

But other CRUD operations require `project_id`, and `search_tool` doesn't return it. So the Agent has to call another tool `get_task_by_id`. On the other hand, MCP-A's query_tasks returns all necessary info to perform the next action in a single call:

Task 1: ID: 6a19143e8f084a8c8101612f Title: buy some groceries Project ID: 6a1914378f084a8c810160a9 Start Date: 2025-07-19 10:00:00 Priority: Medium Status: Active Unfiltered API Data was dumped into context window

If MCP returns pure API results to the Agent's context unprocessed, the Agent's context window will accumulate very fast.

Take MCP-B's `create_task` tool, for example. Its job is to create a to-do item. This is what this tool returns:

{ "id": "6a180de78f086bdead0608be", "projectId": "inbox125587327", ..... "createdTime": "2026-05-28T09:41:59+0000", "modifiedTime": "2026-05-28T09:41:59+0000", "focusSummaries": null }

These 600+ characters mean nothing to the Agent's task, but are still dumped into the Agent's context. On the other hand, MCP-A's create_tasks does a layer of filtering and formatting. This little tweak makes a huge difference in input token usage.

Another issue is tool count. More tools mean a larger candidate set for the model to choose from, which directly increases decision difficulty. In MCP-A, 47 tools were compressed down to 14, covering the same functionality with fewer tools.

---

So here are my takeaways on good MCP tool design: - When designing a tool, think about what the Agent will need next, not just what it's asking for right now. Return enough context in the result so the Agent can take the next action without making another round-trip.

- Too many tools will increase the model's decision burden. So it'd be better to minimize the number of tools within an MCP. Make sure they don't overlap functionalities.

- When your MCP returns data to the LLM, try to keep it LLM-friendly, which means readable. You can filter out unnecessary fields from the API response and format the data, rather than passing through raw JSON.

---

All the tests above were run by MCP-Eval. It's an MCP Server benchmarking tool. If you want to check your MCP's performance, feel free to check this out.

https://github.com/Code-MonkeyZhang/mcp-eval

'World-first' vaccine designed by artificial intelligence

Form design: from zero to hero all in one blog post

CEO to staff: You're not getting a raise. We're spending on AI instead

Bad MCP design cost your Agent 5× more tokens

Delacroix's Entry of the Crusaders into Constantinople Restored

Fedora 43 Upgrade revealed 20 years old Outlook Security Bug

H-1B Crackdown on Indian Workers Erodes a Texas Real Estate Boom

OPRD: On-Policy Representation Distillation

What We Believe Happened: A Summary Timeline of Bricks and Minifigs Store

Verifying Agentic Development at Scale

It's not how much it rose, but how fast

Ask HN: Are people vibe coding internal tools at your companies?

Can These Ads Make You Love A.I.?

Quantum Odyssey

JPMorgan, Citi and Big Banks Plan New Tokenized Deposit System to Answer Crypto

OpenAI's Codex chained decade-old DoS attacks to crash web servers

Bitcoin weathering ugliest week in months as narrative fades, liquidity rotates

Zephir: PHP Extension Language

Show HN: Tatami – a macOS workspace manager with BSP tiling

Aisop – Define AI agent workflows as Mermaid or JSON flow graphs

A Vector Lakebase is all you need for all AI workloads

Reverse Engineering Crazy Taxi part 3

US Tech Sector Announces Most Job Cuts in Nearly Two Years

Benchmarking The BORE Scheduler Performance With CachyOS Linux

Department of Energy Celebrates First Advanced Reactor Criticality

Smart drug strips cancer cells of 'invisibility cloak' can shrink tumours 30%

BPF support in GCC 16 and beyond

RAG Without Persona Modeling Fails Patient Clinical Relevance

What happens if Japan takes in zero immigrants?

Dirk and Linus discuss AI and kernel development

Show HN: Raw Binary Program Analysis Tool