---
id: post/demiliani-com/12455
type: post
title: Model Context Protocol and the “too many tools” problem.
summary: "Summary withheld: an Opus review found that the machine-written summary did not match the post. Read the original post."
tier: community
language: en
tags:
  - mcp
  - context window
  - tool management
  - llm performance
  - agent design
  - api integration
system: copilot
review:
  state: flagged
  by: opus
  at: "2026-10-08T01:59:24.058Z"
  flags: []
generated:
  at: "2026-10-08T01:26:53.141Z"
  pipeline: 0.2.0
  prompts:
    extract-post: 1
  input_hash: 798c1fa28705b8b0e08f6aa78994ab941b8387952f1139a889a69e348706a986
evidence:
  - kind: blog
    url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/
    title: Model Context Protocol and the “too many tools” problem.
    date: "2025-09-04"
    commit: null
    t: null
    quote: null
  - kind: blog
    url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/
    title: Model Context Protocol and the “too many tools” problem.
    date: "2025-09-04"
    commit: null
    t: null
    quote: Having too many tools exposed in an MCP server can degrade the performances and the output of your LLM!
  - kind: blog
    url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/
    title: Model Context Protocol and the “too many tools” problem.
    date: "2025-09-04"
    commit: null
    t: null
    quote: if tools are crowding out the context window, there's less room for context about the current project or goal.
  - kind: blog
    url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/
    title: Model Context Protocol and the “too many tools” problem.
    date: "2025-09-04"
    commit: null
    t: null
    quote: a well-designed MCP server gives agents exactly what they need to accomplish their goals without overwhelming them with choices they don't need.
links:
  learn: []
  objects: []
  features: []
  topics: []
  localizations: []
  videos: []
  posts: []
  guidelines: []
post_id: "12455"
source_id: demiliani-com
source_name: Stefano Demiliani
url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/
published_at: "2025-09-04T08:40:01.000Z"
author: Stefano Demiliani
full_text: false
words: 720
quotes:
  - text: Having too many tools exposed in an MCP server can degrade the performances and the output of your LLM!
    why_it_matters: Directly states the core problem that motivates the entire discussion about tool management
  - text: if tools are crowding out the context window, there's less room for context about the current project or goal.
    why_it_matters: Explains the root cause of performance degradation and why limiting tools matters for agent effectiveness
  - text: a well-designed MCP server gives agents exactly what they need to accomplish their goals without overwhelming them with choices they don't need.
    why_it_matters: Summarizes the design philosophy and best practice recommendation for building effective MCP servers
code_objects_mentioned: []
systems:
  - copilot
  - development
versions_mentioned:
  - claude-sonnet-4
preview:
  embeddable: true
  frame_url: https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/embed/
  image: https://i0.wp.com/demiliani.com/wp-content/uploads/2025/09/mcp_toomanytools_02.png?fit=1200%2C408&ssl=1
  image_alt: null
  image_w: 1200
  image_h: 408
  site_name: Stefano Demiliani
  favicon: https://demiliani.com/wp-content/uploads/2022/09/sd_wordpress_logo.png?w=32
  probed_at: "2026-10-07T11:51:39.561Z"
---

# Model Context Protocol and the “too many tools” problem.

[Read the post](https://demiliani.com/2025/09/04/model-context-protocol-and-the-too-many-tools-problem/) · Stefano Demiliani (Stefano Demiliani, MVP) · 2025-09-04 · 720 words · tier community · **flagged** (a review found a problem)

> Summary withheld: an Opus review found that the machine-written summary did not match the post. Read the original post.

## Quotes

- "Having too many tools exposed in an MCP server can degrade the performances and the output of your LLM!" (Directly states the core problem that motivates the entire discussion about tool management)
- "if tools are crowding out the context window, there's less room for context about the current project or goal." (Explains the root cause of performance degradation and why limiting tools matters for agent effectiveness)
- "a well-designed MCP server gives agents exactly what they need to accomplish their goals without overwhelming them with choices they don't need." (Summarizes the design philosophy and best practice recommendation for building effective MCP servers)

## Context

- Features: MCP server autodiscovery, Model Context Protocol, Claude Code, Cursor MCP integration, GitHub Copilot MCP support, MCP Proxy
- Versions: claude-sonnet-4

Source: Stefano Demiliani, community blog. Summary, key points and quotes are derived (CONTENT-NOTICE.md); read the original for the full text.