de en es ja ko ru zh

MetaMCP

Name: MetaMCP
Author: mentu-ai

Official

by mentu-ai

Overview Schema Related Servers Score Discussions

TypeScript

Hybrid

MetaMCP

npm version Node.js License

MetaMCP ist ein sicheres, bedarfsgesteuertes Gateway für den langen Schwanz von MCP-Servern. Es bietet einem MCP-Client drei stabile Tools:

mcp_discover findet konfigurierte Server, gecachte Tool-Schemas und überprüfte Methoden, ohne jedes Kind zu starten.
mcp_call ruft lazy ein explizit benanntes Kind-Tool auf.
mcp_run führt eine begrenzte, schema-validierte deklarative Methode aus.

MetaMCP ist bewusst kein Ersatz für jede direkte MCP-Verbindung. Behalten Sie wichtige, häufig genutzte, kompakte oder stark authentifizierte MCPs direkt. Legen Sie unregelmäßige Long-Tail-Server hinter MetaMCP und fördern Sie wiederholte Multi-Schritt-Rituale zu Methoden.

                               ┌─ direct: GitHub / Codex Apps / core runtime
MCP client ────────────────────┤
                               └─ MetaMCP (3 tools)
                                    ├─ discover cached capabilities
                                    ├─ call one lazy child
                                    └─ run reviewed Methods

Wann welcher Pfad verwendet werden sollte

Pfad	Beste Eignung	Warum
Direktes MCP	Hochfrequente, kompakte, sicherheitskritische oder grundlegende Server	Bewahrt typisierte Schemas, native Authentifizierung und explizite Genehmigungen
`mcp_discover` + `mcp_call`	Long-Tail- oder unregelmäßige Fähigkeiten	Hält die Client-Oberfläche klein, ohne die Assemblersprache zu verbergen
`mcp_run`	Wiederholte Acquire → Normalize → Analyze-Workflows	Macht begrenztes Verhalten testbar, versioniert und evidenzproduzierend

Leiten Sie Abrechnung, Infrastrukturmutation, Identität oder einen anderen Server mit hohen Konsequenzen nicht nur zur Reduzierung der Tool-Anzahl durch MetaMCP. Die richtige Grenze ist operativ, nicht ideologisch.

Related MCP server: mcp-gateway

Schnellstart

Erfordert Node.js 20 oder neuer.

npx @mentu/metamcp@latest --config .mcp.json

Untersuchen Sie die vollständige modellseitige Oberfläche, bevor Sie einen Client konfigurieren:

npx @mentu/metamcp@latest tools
npx @mentu/metamcp@latest tools --json

Der Inspector liest dieselben Definitionen, die von MCP tools/list zurückgegeben werden, und beendet sich dann, bevor er Konfiguration lädt, Speicher öffnet, ein Kind startet oder einen Transport bindet. --json enthält die vollständigen Eingabeschemas für automatisierte Überprüfung und Versionsunterschiede.

Erstellen Sie .mcp.json:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem@2026.7.10", "/path/to/allowed/files"]
    },
    "internal-api": {
      "command": "node",
      "args": ["./servers/internal-api.js"],
      "env": { "API_TOKEN": "${INTERNAL_API_TOKEN}" },
      "inheritEnv": ["HTTP_PROXY"]
    }
  }
}

Kind-Server starten nur, wenn sie explizit aktualisiert, aufgerufen oder von einer Methode verwendet werden. Einfache Erkennung liest Konfiguration und gecachte Schemas; sie startet nicht alle Kinder.

Servernamen sind stabile Cache-Identitäten und müssen 1–128 Buchstaben, Zahlen, Punkte, Unterstriche oder Bindestriche enthalten; Pfadtrennzeichen und traversal-ähnliche Namen werden abgelehnt.

Sichere Client-Einrichtung

init ist nur Vorschau, es sei denn, --yes wird angegeben. Ohne einen benannten Client werden nur vorhandene Client-Konfigurationsdateien berücksichtigt.

metamcp init                         # preview, no writes
metamcp init --client Codex          # preview one client
metamcp init --client Codex --yes    # apply atomically and write a .bak

Fehlerhaftes JSON wird abgelehnt und unverändert gelassen. Ein benannter Client kann explizit erstellt werden; MetaMCP erstellt niemals standardmäßig jede unterstützte Client-Konfiguration.

Verwenden Sie für eine manuelle Client-Konfiguration einen absoluten Pfad, damit das Gateway nicht vom Arbeitsverzeichnis des Clients abhängt:

{
  "mcpServers": {
    "metamcp": {
      "command": "npx",
      "args": [
        "-y",
        "@mentu/metamcp@latest",
        "--config",
        "/absolute/path/to/.mcp.json"
      ]
    }
  }
}

@latest ist für die Evaluierung praktisch. Pinnen Sie @mentu/metamcp@1.0.0 in kontrollierten Umgebungen, damit Upgrades bewusst und überprüfbar sind.

Die drei Tools

Discover

{ "query": "capture screenshot", "kind": "tool" }

Die Erkennung durchsucht nur Live- oder gecachte Schemas. Um einen Server aus seiner Live-Tool-Liste zu aktualisieren:

{ "server": "browser", "refresh": true }

refresh ohne Server wird abgelehnt, damit ein Agent nicht versehentlich über die gesamte Konfiguration ausbreitet.

Call

{
  "server": "browser",
  "tool": "capture_page",
  "args": { "url": "https://example.com" },
  "timeoutMs": 60000
}

MetaMCP spielt einen Kind-Aufruf nach einem Timeout oder Transportfehler niemals automatisch erneut ab. Das Kind könnte eine Mutation abgeschlossen haben, bevor die Antwort verloren ging. Eine spätere Methode darf nur dann erneut versuchen, wenn ihr Manifest diesen Schritt explizit als idempotency: "safe" deklariert.

Eine Methode ausführen

Legen Sie JSON-Manifeste in .metamcp/methods/ ab oder übergeben Sie --methods <Verzeichnis>. Die Kind-Server- und Tool-Namen unten sind illustrativ; binden Sie sie an überprüfte Server in Ihrer eigenen Konfiguration:

{
  "apiVersion": "metamcp.io/v1alpha1",
  "kind": "Method",
  "metadata": {
    "name": "content.acquire-and-normalize",
    "version": "1.0.0",
    "description": "Acquire content and normalize it into a stable record"
  },
  "spec": {
    "effects": "read",
    "inputSchema": {
      "type": "object",
      "properties": { "url": { "type": "string" } },
      "required": ["url"],
      "additionalProperties": false
    },
    "steps": [
      {
        "id": "acquire",
        "server": "fetch",
        "tool": "fetch",
        "args": { "url": "${input.url}" }
      },
      {
        "id": "normalize",
        "server": "content",
        "tool": "normalize",
        "dependsOn": ["acquire"],
        "args": { "document": "${steps.acquire.structuredContent}" }
      }
    ],
    "output": "${steps.normalize.structuredContent}"
  }
}

Dann rufen Sie auf:

{ "method": "content.acquire-and-normalize", "input": { "url": "https://example.com" } }

Methoden sind deklarativ und nicht beliebiges JavaScript. Sie haben begrenzte Schrittzahlen, Fristen und Ausgabegrößen; Eingabe-/Ausgabe-JSON-Schemas; explizite Lese-/Schreibeffekte; sichere Interpolation; typisierte Lücken; und eine Schritt-für-Schritt-Ablaufverfolgung. Schreib- oder gemischte Effekt-Methoden sind deaktiviert, es sei denn, der Gateway-Betreiber startet MetaMCP mit --allow-writes.

Siehe Method Mode, das Manifest-Schema und die Beispielmethode. Das Design verallgemeinert die Konsistenzschicht, die von Crawlio Method Mode dokumentiert wird.

Konfiguration und Geheimnisse

${NAME}-Referenzen in env und HTTP-headers werden standardmäßig aus der Host-Umgebung aufgelöst. Eine unaufgelöste Referenz führt zu einem Startfehler; sie wird niemals als wörtlicher Platzhalter an ein Kind übergeben.

MetaMCP kopiert seine Umgebung nicht in stdio-Kinder. Es erbt nur eine kleine Laufzeit-Allowlist (PATH, Home-/Temp-/Locale-Variablen und Plattformäquivalente), Variablen, die in inheritEnv benannt sind, und Werte, die explizit im Kind-env-Block gesetzt sind. Einbettende können einen benutzerdefinierten SecretProvider für einen Schlüsselbund oder Tresor installieren.

Die Erkennung ist standardmäßig eine lokale Stichwortsuche. Um sich für die Voyage-gestützte semantische Suche zu entscheiden, setzen Sie METAMCP_VOYAGE_API_KEY explizit; Erkennungsabfragen werden dann an Voyage gesendet und der optionale lokale SQLite-Vektorindex wird aktiviert. Umgebungsvariablen wie ANTHROPIC_API_KEY oder VOYAGE_API_KEY aktivieren niemals Netzwerkaufrufe.

Remote-Kind-Server verwenden url, transportType, headers und die vorhandenen OAuth-Felder:

{
  "mcpServers": {
    "remote": {
      "url": "https://mcp.example.com/mcp",
      "transportType": "http",
      "headers": { "Authorization": "Bearer ${REMOTE_TOKEN}" }
    }
  }
}

HTTP-Gateway

Der HTTP-Modus bindet standardmäßig an 127.0.0.1:

metamcp --transport http --port 8080 --config .mcp.json

Eine nicht authentifizierte Nicht-Loopback-Bindung schlägt geschlossen fehl. Konfigurieren Sie die OAuth-Ressourcenserver-Validierung oder METAMCP_HTTP_BEARER_TOKEN, bevor Sie den Listener freigeben. Browseranfragen mit einem Origin-Header werden abgelehnt, es sei denn, die genaue Herkunft wird mit --allow-origin oder METAMCP_ALLOWED_ORIGINS angegeben.

MetaMCP bedient Legacy-MCP-Clients und das zustandslose Anfrage-Envelope von 2026-07-28 über stdio und Streamable HTTP. Siehe Architektur für die unterstützte Grenze und Bereitstellungsanleitung.

Evidenz

Abgeschlossene mcp_call- und mcp_run-Versuche werden in .metamcp/ledger.jsonl serialisiert. Exportieren Sie ein portables, hash-verknüpftes Bündel:

metamcp export-evidence \
  --ledger .metamcp/ledger.jsonl \
  --out .metamcp/evidence-bundle.json

metamcp export-evidence --out .metamcp/evidence-bundle.json --verify

Das operative Ledger ist kein Remote-Attestierungssystem. Der Export erkennt spätere Änderungen innerhalb eines Bündels; er beweist nicht, dass ein kompromittierter Host jedes Ereignis aufgezeichnet hat.

Optionale Galerie

Das Paket enthält weiterhin eine menschlich betriebene Server-Galerie:

metamcp add --list
metamcp add playwright sentry --config .mcp.json

Die Laufzeit installiert niemals Pakete als Reaktion auf einen MCP-Tool-Aufruf. Die Installation bleibt eine explizite CLI-/Benutzeraktion.

Upgrade von 0.x

Version 1.0 entfernt absichtlich die modellseitigen Provisionierungs-, Skill-Beratungs- und JavaScript-Ausführungstools. Sie ändert auch HTTP-Bindung, Kind-Umgebungsvererbung, Wiederholungen und init. Lesen Sie Migration zu 1.0 vor dem Upgrade.

Sicherheit

Kind-MCP-Server sind vertrauenswürdiger lokaler oder Remote-Code mit eigenen Berechtigungen. MetaMCP ist eine Richtlinien- und Lebenszyklusgrenze, keine OS-Sandbox für nicht vertrauenswürdige Pakete. Überprüfen Sie Befehle, pinnen Sie Pakete wo angemessen, begrenzen Sie Anmeldeinformationen pro Kind und halten Sie gefährliche direkte Server hinter clientseitiger menschlicher Genehmigung.

Melden Sie Schwachstellen privat, wie in SECURITY.md beschrieben.

Entwicklung

npm ci
npm run typecheck
npm test
./scripts/smoke-test.sh
npm run check:release
npm pack --dry-run

npm publish führt erneut die vollständige verify:release-Gate aus. Das Gate leitet die öffentliche Tool-Oberfläche aus der gebauten CLI ab und prüft Paket, Lockfile, Changelog und offizielle MCP-Registry-Metadaten auf Versionsabweichungen.

Apache-2.0-lizenziert. Gepflegt von Mentu AI.

Available Tools

6 tools

mcp_callB

Forward a tool call to a specific child MCP server. Retries once on crash.

ParametersJSON Schema

Name	Required	Description
`args`	No	Arguments to pass to the tool
`tool`	Yes	Tool name to call
`server`	Yes	Target server name

TDQS

B3.2/5.0

Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Discloses retry behavior once on crash, which is useful. But with no annotations, the description does not mention side effects, permissions, or whether the operation is destructive. The retry detail is positive but incomplete.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences with no waste. The main action and key behavior (retry) are front-loaded. Every sentence serves a purpose.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With no output schema and subjective complexity medium, the description omits important context: what happens if the server is unreachable, what the return value is, and any rate limits or error handling details. Only minimal forwarding and retry are covered.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the schema already describes all parameters. The description adds no additional meaning beyond the schema. Baseline 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool forwards a tool call to a child MCP server and retries on crash. It uses a specific verb and resource, but does not explicitly contrast with sibling tools like mcp_execute or mcp_discover.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance on when to use this tool versus alternatives. The description lacks context about prerequisites, alternatives, or scenarios where this tool is preferred.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

mcp_discoverA

Search tool catalogs across all child MCP servers + list server status. If no query, returns server list with status and tool counts.

ParametersJSON Schema

Name	Required	Description	Default
`query`	No	Search query for tools
`server`	No	Filter to a specific server

TDQS

A4/5.0

Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It describes the tool as performing read operations (search and list), but does not disclose any behavioral traits like authentication requirements, rate limits, or potential side effects. It is adequate but lacks depth.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is two sentences, concise, and front-loaded with the core purpose. Every sentence adds value, with no wasted words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

The description covers both modes of the tool (search and list server status) and explains what happens when no query is provided. No output schema is given, but the description implies the return structure sufficiently.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% with both parameters described. The description adds some context by explaining the no-query behavior, but does not substantially enhance the parameter definitions beyond what the schema provides.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool searches tool catalogs across child MCP servers and lists server status. It uses the verb 'discover' and distinguishes from siblings like mcp_call (which likely calls a tool) and mcp_skill_discover (which discovers skills).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explains two usage modes: with a query (search) and without a query (list server status). It provides context on when to use each, though it does not explicitly say when not to use or mention alternatives.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

mcp_executeA

Code-mode execution in V8 sandbox. Access provisioned servers via servers.<name>.call(tool, args). Supports async/await, sleep(ms), console.log.

ParametersJSON Schema

Name	Required	Description	Default
`code`	Yes	Code to execute

TDQS

A3.8/5.0

Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations provided, so description shoulders full burden. It discloses sandboxing, async/await, sleep, and server access, but omits details like error handling, return value format, persistence, or side effects.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Two sentences, front-loaded with purpose, every sentence adds value. No redundant or extraneous information.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Lacks critical details for a code execution tool: what is returned (execution result), restrictions (network, filesystem), lifecycle (one-shot), and output capture. Relies heavily on inferred context from sibling tools.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema only describes 'code' as 'Code to execute'. Description adds significant context: V8 sandbox, ability to use async/await, sleep, console.log, and call provisioned servers. This goes beyond the bare schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Description clearly states it executes code in a V8 sandbox, with explicit mention of accessing provisioned servers. This distinguishes it from sibling tools like mcp_call which directly call server tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies usage for running arbitrary JavaScript with sandbox access, and contrasts with direct server calls by showing servers.<name>.call syntax. However, it lacks explicit when-to-use vs. alternatives guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

mcp_provisionB

Intent-based provisioning. Describe what you need, MetaMCP resolves and provisions the right server.

ParametersJSON Schema

Name	Required	Description
`intent`	Yes	What capability you need
`context`	No	Additional context for resolution
`autoProvision`	No	Auto-provision if trusted (default: false)

TDQS

B3.3/5.0

Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations provided, so description must fully explain behavior. It mentions intent-based provisioning but lacks details on side effects (e.g., resource creation, persistence), permissions required, or potential destructive actions.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Single sentence is concise and front-loaded with core purpose. However, for a provisioning tool, it may be too terse, lacking necessary detail.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With only 3 simple parameters and no output schema, the description is too brief for a provisioning action. It does not describe return values, success indicators, or consequences, leaving the agent to infer too much.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3. Description adds no extra meaning beyond the schema descriptions; no examples or clarifications on intent format.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Description clearly states it is for provisioning servers based on intent ('Describe what you need, MetaMCP resolves and provisions the right server'). Distinct from siblings like mcp_discover or mcp_call which focus on discovery or execution.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No explicit guidance on when to use vs alternatives. Usage is implied: for provisioning when you know the desired capability. Lacks exclusion criteria or prerequisites.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

mcp_skill_adviseA

Pre-flight readiness check for a skill. Returns MCP server availability and recommendations.

ParametersJSON Schema

Name	Required	Description	Default
`skill`	Yes	Skill name to check

TDQS

A3.8/5.0

Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the burden. It reveals that the tool returns 'MCP server availability and recommendations', which is useful. However, it does not disclose edge cases (e.g., skill not found), permissions needed, or potential side effects.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single concise sentence that effectively conveys the purpose and output without unnecessary words.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the simplicity of the tool (1 param, no output schema), the description is mostly adequate. It explains the return value, but could be more complete by mentioning that it is non-destructive or clarifying what 'recommendations' entails.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% for the single parameter 'skill'. The description adds no extra meaning beyond what the schema already provides, meeting the baseline of 3.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states it is a 'Pre-flight readiness check for a skill', specifying the verb and resource. It distinguishes from siblings like mcp_skill_discover (discovery) and mcp_call (execution) by focusing on readiness checking.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies use before calling or executing a skill, but lacks explicit when-to-use or when-not-to-use guidance. No alternatives are mentioned despite having several sibling tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

mcp_skill_discoverA

Search Claude Code skills with MCP readiness status. Returns skills matching query with their required MCP servers and availability.

ParametersJSON Schema

Name	Required	Description	Default
`query`	Yes	Search query for skills
`domain`	No	Filter by domain (e.g. browser_automation, monitoring)

TDQS

A4/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden and discloses that results include required MCP servers and availability—key behavioral traits. It does not mention auth or rate limits, which is acceptable for a read-only search tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, efficient sentence that immediately conveys the core purpose and key output details.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a search tool with no output schema, the description provides reasonable completeness by listing what is returned (skills, required servers, availability). It could be improved by noting pagination or result limits.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and the schema descriptions for 'query' and 'domain' are adequate. The tool description adds no additional meaning beyond the schema, so a baseline of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool searches Claude Code skills, specifies the 'MCP readiness status' criterion, and indicates the return includes required MCP servers and availability. This distinguishes it from sibling tools like mcp_skill_advise and mcp_discover.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies the tool should be used when searching for skills with MCP readiness, but provides no explicit guidance on when not to use it or how it compares to other search/discovery tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

6 tool updatesv0.5.0Jul 28, 2026
- First observedmcp_call
- First observedmcp_discover
- First observedmcp_execute
- First observedmcp_provision
- First observedmcp_skill_advise
- First observedmcp_skill_discover

Install Server

RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

A3.9/5.0

Disambiguation5/5

Each tool has a clearly distinct purpose: discovery of servers/tools, provisioning, forwarding tool calls, code execution, skill discovery, and skill readiness checking. No two tools have overlapping functionality.

Naming Consistency5/5

All tools follow the 'mcp_<verb>' pattern, with skill-specific tools adding 'skill_' for clarity. The naming is consistent and predictable.

Tool Count5/5

With 6 tools, the set is well-scoped for a meta-server. It covers the essential operations without being excessive or insufficient.

Completeness4/5

The tool surface covers discovery, provisioning, calling, execution, and skill support. Minor gaps exist (e.g., no explicit unprovisioning or server management), but core agent workflows are well-supported.

Maintenance

ActivityMaintained

ResponsivenessUnresponsive

Resources

Need Help?

Related Servers

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

ToolRouter B
io.github.Humanleap
The OpenRouter for tools. One MCP connection gives any AI agent 254 hosted tools, pay per call.
Agent Orchestration Autonomous Agents
471
Unyly GatewayOAuth
org.unyly
One connector for 15,000+ MCP servers plus your team's private MCPs, from any AI client.
Agent Orchestration AI & Machine Learning
2
MCP Aggregation Gateway — 22k+ MCP search & call A
cn.pianam.mcp
One AI endpoint to search and call 22k+ MCP servers; 50+ hosted tools work instantly, no key.
Agent Orchestration Search
31
Artificial Wit AIOAuth
com.artificialwit
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
ERP Systems CRM Knowledge & Memory
2

Related MCP Servers

MCP Gateway
Developer Tools Agent Orchestration App Automation
LunFengChen
F
license
A
quality
C
maintenance
Aggregates multiple MCP servers into a single gateway with unified top-level tools, reducing LLM context usage and enabling IDE compatibility by consolidating many tools into fewer interface functions.
Updated 2 months ago (2026-07-12 14:16 UTC)
4
17
-
mcp-gateway
Developer Tools App Automation
jj13jp
F
license
Not graded
quality
B
maintenance
Aggregates multiple child MCP servers into a single MCP server endpoint, enabling clients to use various tools (e.g., filesystem, Brave Search) through one interface.
Updated 3 months ago (2026-06-19 11:39 UTC)
19
-
Agent Aggregator
Agent Orchestration AI & Machine Learning
rnd-pro
A
license
A
quality
D
maintenance
Aggregates tools from multiple MCP servers, acting as a proxy to provide unified access to various AI agents and tools.
Updated 3 months ago (2026-06-03 17:27 UTC)
10
15
3
MIT
MCP Server Proxy
Developer Tools Agent Orchestration
hr98w
A
license
Not graded
quality
D
maintenance
Aggregates multiple MCP servers into a single endpoint, enabling LLM clients to access tools, resources, and prompts from various backends through one connection.
Updated a year ago (2025-09-11 05:39 UTC)
18
MIT

Latest Blog Posts

Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)
By Om-Shree-0709 on July 25, 2026.
mcp
Agent Identity
OAuth 2.1
Your AI Chatbot Just Exposed Your CEO's Salary to an Intern
By Om-Shree-0709 on July 2, 2026.
Agent Identity
MCP Security
OAuth Delegation
Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)
By Om-Shree-0709 on June 30, 2026.
Agentic Ai
Prompt Injection
WebAssembly

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/mentu-ai/metamcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server