Jump To Category Page

Jump To Representative Detail

MCP Servers Category View

MCP server inventory should show which tool surfaces widen reach, how broadly they are shared, and where health or labeling gaps change trust.

The category page should treat each server as a capability surface shared across sessions rather than as a flat integration list.

Use This Page When Asking

Which tool surfaces widen reach enough that they need focused scrutiny?

Representative Detail

filesystem@1.0.0

Open representative detail

Servers observed

Mixed transports

Tool surface

32 tools

Reach concentrated in a few servers

Shared sessions

Broad caller reuse matters most

Labeling gaps

1 new surface

Weak business context

Why This Page Exists

This category page should help the reviewer decide what matters in mcp servers before the raw table takes over the reading experience.

What This Page Should Make Obvious

The reviewer should know what is routine, what is risky, and what deserves the next click.

Treat servers as access surfaces

A widely shared server matters because of what it lets agents reach and how broadly that reach is reused.

Operational health changes trust

Latency, error rate, or unlabeled transport should be visible as first-order interpretation signals.

Keep top tools visible

A server is easier to judge when the page shows its real effective surface, not just a name and version.

Fastest Drill Paths

The next clicks should feel obvious and intentional.

Shared filesystem lane

One server shapes the majority of code-oriented sessions and file reach.

Open representative detail

It has the broadest caller demand and the clearest consequence.

Raw transport and error table

Reviewers still need the exact server inventory after the interpretation layer.

Open c6 raw list

The raw list is the follow-up check for precise counts and error rates.

MCP Server-Centered Topology

An MCP server map should expose caller demand, tool surface, transport, and operational health.

Useful per server: which agents connect to it, how many tools it exposes, what sessions depend on it, where transport or latency adds risk, and what failures cluster around it.

MCP Server

filesystem v1.0.0

Broadly shared server whose tools shape the majority of code-oriented sessions.

18 sessions · 420 calls

Transport

http

Tool surface

8 tools

Avg posture

0.2% error rate

Blast radius

shared by multiple agents

Callers

Who depends on this server.

3 nodes

Claude Codeprimary agent

Cursorsecondary agent

alice@company.comfrequent user

Tool Surface

Which exposed tools matter most.

3 nodes

read_file318 calls

write_file204 calls

list_directory191 calls

Execution Context

Operational details that explain behavior.

3 nodes

http transportnetworked

local filesystemasset reach

todaylast seen

Watch For

Where health and risk can degrade.

3 nodes

web-search 3.4%compare error outlier

latency drifttool slowdown

unexpected callernew client

c6 Raw Reference

c6 shape: server catalog with transport, tool counts, sessions, and errors.

1 server path needs labeling

A newly seen MCP surface has tool reach but weak business-purpose context.

Servers

Tools exposed

Connected sessions

Highest-error server

filesystem@1.0.0

Why Keep The Raw Ledger

c12 should still flow back into the raw inventory model. This page frames the question and likely answer; the raw table proves it.

Open c6 Raw List

Use the original category ledger for exact rows, counts, and timestamps.

Open c6 raw list

Raw MCP Server Inventory

Server, transport, tools, sessions, calls, and error rate.

Server	Transport	Tools	Sessions	Error rate
filesystem@1.0.0	stdio	8	6	2.3%
github@1.4.0	http	12	3	0.0%
browser@0.9.2	stdio	5	2	1.1%