Agent skill

Mcpc Core

by mcpc-tech in mcpc-tech/mcpc

Build agentic MCP servers using @mcpc/core by composing existing MCP tools.

MITAuto-check passedAgent Workflows

Install Mcpc Core

skills CLI
$ npx skills add mcpc-tech/mcpc --skill mcpc-core -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install mcpc-tech/mcpc mcpc-core --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/mcpc-tech/mcpc.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/mcpc-core .claude/skills/mcpc-core && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
mcpc-core
GitHub stars
103
Token cost
~2.2k tokens
SKILL.md length
320 words
Files
2 (incl. references)
Skills in repo
4
Repo updated
First seen
Licence
MIT

At a glance

Build agentic MCP servers using @mcpc/core by composing existing MCP tools.

  • Debugging MCPC agentic servers — including setting up compositions
  • SKILL.md covers Installation, Minimal Example, Core API and Tool References ( XML), plus 8 more sections
  • Calls deno, npm and pnpm
  • Choosing execution modes

What it does

Mcpc Core is an agent skill from mcpc-tech/mcpc. Build agentic MCP servers using @mcpc/core by composing existing MCP tools. Use when creating, extending, or debugging MCPC agentic servers — including setting up compositions, choosing execution modes, writing plugins, loading skills, and connecting transports.

Its SKILL.md is about 2.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files, including reference files (for example `references/examples.md`).

It sits in Agent Workflows, covering MCP servers. It works with Model Context Protocol. The repository describes itself as: Build agentic-MCP servers by composing existing MCP tools. The licence is MIT.

When your agent uses it

  • Debugging MCPC agentic servers — including setting up compositions
  • Choosing execution modes
  • Writing plugins
  • Connecting transports

Example prompts

  • “/mcpc-core”

Requirements

  • Node.js

What it can do on your machine

Read from SKILL.md and the folder at commit 918ffef. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • deno
    • npm
    • pnpm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm and pnpm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Mcpc Core loads about 2.2k tokens when it runs, and up to ~3.7k if it reads all its reference files. Until then it costs about 68 tokens; SKILL.md has 320 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~68
When it runs · the whole SKILL.md, loaded when a task matches
~2.2k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from mcpc-tech/mcpc at commit 918ffef, republished under its MIT licence (© mcpc-tech). 320 words, ~2,184 tokens.

Download SKILL.mdSave it as .claude/skills/mcpc-core/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
mcpc-core
description
Build agentic MCP servers using @mcpc/core by composing existing MCP tools. Use when creating, extending, or debugging MCPC agentic servers — including setting up compositions, choosing execution modes, writing plugins, loading skills, and connecting transports.

@mcpc/core — Build Agentic MCP Servers

@mcpc/core lets you compose existing MCP tools into agentic MCP tools. You write a description that references tools via <tool> XML tags, and MCPC wires everything up into a working MCP server.

Installation

bash
# Deno (preferred in this repo)
deno add jsr:@mcpc/core

# npm
npm install @mcpc-tech/core

# pnpm (from JSR)
pnpm add jsr:@mcpc/core

Minimal Example

typescript
import { mcpc } from "@mcpc/core";
import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";

const server = await mcpc(
  [{ name: "my-agent", version: "1.0.0" }, { capabilities: { tools: {} } }],
  [{
    name: "file-agent",
    description: `I help manage files.
Available tools:
<tool name="@wonderwhy-er/desktop-commander.read_file"/>
<tool name="@wonderwhy-er/desktop-commander.write_file"/>`,
    deps: {
      mcpServers: {
        "@wonderwhy-er/desktop-commander": {
          command: "npx",
          args: ["-y", "@wonderwhy-er/desktop-commander@latest"],
          transportType: "stdio",
        },
      },
    },
  }],
);

await server.connect(new StdioServerTransport());

Core API

mcpc(serverConf, composeConf?, options?)
ParamTypeDescription
serverConf[ServerMeta, ServerCapabilities]Name/version + capabilities tuple
composeConfComposeInput[]Array of ComposeDefinition objects or .md file paths
optionsMcpcOptionsLoader plugins + setup callback

Returns Promise<ComposableMCPServer>.

ComposeDefinition
typescript
{
  name: string | null,          // null = composition-only (no tool exposed)
  description?: string,         // Natural language + <tool> XML refs
  manual?: string,              // Progressive disclosure: long docs hidden behind man tool
  deps?: MCPSetting,            // Dependent MCP servers to connect
  plugins?: (ToolPlugin | string)[],  // Runtime plugins
  options?: {
    mode?: "agentic" | "ai_sampling" | "ai_acp",  // default: "agentic"
    maxSteps?: number,          // default: 50
    maxTokens?: number,         // default: 128_000 (ai_sampling only)
    samplingConfig?: SamplingConfig,
    providerOptions?: { modelPreferences?: {...} },  // ai_sampling only
    acpSettings?: { command, args, env, session, persistSession },  // ai_acp only
    tracingEnabled?: boolean,   // OpenTelemetry
    refs?: ToolRefXml[],
  }
}

Tool References (<tool> XML)

In the description field, reference tools with XML tags:

<tool name="server.tool_name"/>              — specific tool
<tool name="server.__ALL__"/>               — all tools from a server
<tool name="server.tool" maxResultLength="2000"/>  — truncate large results
<tool name="server.tool" hide/>             — use internally, hide from public API
<tool name="server.tool" global/>           — expose at top-level scope

MCP Server Dependencies (deps.mcpServers)

Three transport types:

typescript
// stdio
"my-server": { command: "npx", args: ["-y", "pkg@latest"], transportType: "stdio" }

// Streamable HTTP
"remote": { transportType: "streamable-http", url: "https://api.example.com/mcp/",
            headers: { "Authorization": "Bearer ${TOKEN}" } }

// SSE
"sse-server": { transportType: "sse", url: "https://api.example.com/sse/" }

Execution Modes

ModeDescriptionBest For
agenticFully autonomous, uses MCP samplingDefault; most flexible
ai_samplingUses AI SDK streamText + MCP sampling providerWhen you control the model
ai_acpAI SDK + ACP protocol for coding agentsClaude Code, Codex-style agents

Plugins

Built-in Plugins (import from @mcpc/core/plugins)
typescript
import {
  createBashPlugin, // Add bash execution tool
  createLargeResultPlugin, // Truncate/paginate large tool results
  createSearchPlugin, // Add ripgrep-based search tool
  createSkillsPlugin, // Lazy-load skill knowledge
} from "@mcpc/core/plugins";
Using Plugins
typescript
// In ComposeDefinition
plugins: [
  createLargeResultPlugin({ maxSize: 8000 }),
  createSkillsPlugin({ paths: [".agents/skills"] }),
  createBashPlugin(),
  "./my-plugin.ts?param=value", // File path with query params
];

// Or in mcpc() options (loader plugins, run before composition)
await mcpc(serverConf, composeConf, {
  plugins: ["@mcpc/plugin-markdown-loader"],
});
Writing a Custom Plugin (ToolPlugin)
typescript
import type { ToolPlugin } from "@mcpc/core";

const myPlugin: ToolPlugin = {
  name: "my-plugin",

  // Mutate/filter tool descriptions at compose time
  transformTool?: (tool: ComposedTool, ctx: TransformContext) => ComposedTool,

  // Intercept inputs before execution
  transformInput?: (args: unknown, ctx: BeforeToolExecuteContext) => unknown,

  // Intercept outputs after execution
  transformOutput?: (result: unknown, ctx: AfterToolExecuteContext) => unknown,

  // Lifecycle hooks
  onComposeStart?: (ctx: ComposeStartContext) => void,
  onComposeEnd?: (ctx: ComposeEndContext) => void,
  onFinalize?: (ctx: FinalizeContext) => void,
};

Skills Plugin

createSkillsPlugin adds a {agent}__load-skill tool that lazy-loads skill content:

typescript
createSkillsPlugin({
  paths: [join(projectRoot, ".agents/skills")],
});

Skills directory structure:

.agents/skills/
└── my-skill/
    ├── SKILL.md          ← main instructions (loaded by default)
    └── references/
        └── api.md        ← loaded via { skill, ref: "references/api.md" }

Agent calls:

agent__load-skill({ skill: "my-skill" })
agent__load-skill({ skill: "my-skill", ref: "references/api.md" })

Markdown Agent Files

Load agent definitions from .md files (requires @mcpc/plugin-markdown-loader):

typescript
await mcpc(
  [{ name: "server", version: "1.0.0" }, { capabilities: { tools: {} } }],
  ["./agents/my-agent.md"],
  { plugins: ["@mcpc/plugin-markdown-loader"] },
);

Markdown format:

markdown
---
name: my-agent
deps:
  mcpServers:
    some-server:
      command: npx
      args: ["-y", "some-server"]
      transportType: stdio
options:
  mode: agentic
---

I am an agent that does X.

<tool name="some-server.tool"/>

Setup Callback (Custom Tools)

Register tools directly on the server before composition:

typescript
import { jsonSchema } from "@mcpc/core";

await mcpc(serverConf, composeConf, {
  setup: async (server) => {
    server.tool(
      "my-tool",
      "Description of tool",
      jsonSchema({ type: "object", properties: { input: { type: "string" } } }),
      async (args) => ({
        content: [{ type: "text", text: `Got: ${args.input}` }],
      }),
    );
    await server.addPlugin(myPlugin);
  },
});

Gotchas

  • transportType is required in every mcpServers entry — omitting it causes a silent failure.
  • name: null in ComposeDefinition skips creating the composed tool; useful for composition-only (just wiring up deps + plugins).
  • Plugin load order matters: McpcOptions.plugins run before composition (use for loaders). ComposeDefinition.plugins run after composition (use for runtime transforms).
  • hide vs global on <tool>: hide keeps a tool internal to the agent. global exposes it at top-level without a namespace prefix.
  • manual field: if set, description becomes a short summary and full content is behind a man tool call — reduces initial context size.
  • Environment variables in headers: use ${VAR_NAME} syntax inside strings in deps.mcpServers[*].headers — MCPC expands them at runtime.
  • maxSteps default is 50 for ai_sampling/ai_acp modes — increase for complex multi-step tasks.

References

© mcpc-tech, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in .agents/skills/mcpc-core of mcpc-tech/mcpc.

  • SKILL.md
  • references/examples.md

Open the folder on GitHubat commit 918ffef

Compare with similar skills

Mcpc Core next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Mcpc Core compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Mcpc Core this skillmcpc-tech/mcpc103—~2.2kAutomated safety check: PassMIT
MCP Server Builderanthropics/skills180k64 repos~2.3kAutomated safety check: PassApache-2.0
MCP Server BuildershareAI-lab/learn-claude-code78k5 repos~1.2kAutomated safety check: PassMIT
MCP Integration for Pluginsanthropics/claude-plugins-official38k11 repos~3.1kAutomated safety check: PassApache-2.0
Fastmcp Client CLIPrefectHQ/fastmcp28k1 repos~823Automated safety check: PassApache-2.0
Crush Configurationcharmbracelet/crush29k—~3.7kAutomated safety check: PassCustom licence

Similar skills

  • MCP Server Builder

    anthropics/skills

    Official

    Guides the design and implementation of Model Context Protocol servers in TypeScript or Python, from tool naming and error messages to evaluation.

    180k GitHub starsUsed in 64 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • MCP Server Builder

    shareAI-lab/learn-claude-code

    Walks through building MCP servers in Python or TypeScript that expose tools, resources and prompts to Claude, with templates, registration and testing.

    78k GitHub starsUsed in 5 repos~1.2k tokens
    Agent WorkflowsAuto-check passed
  • MCP Integration for Plugins

    anthropics/claude-plugins-official

    Official

    Explains how to bundle Model Context Protocol servers in a Claude Code plugin, covering config files, stdio, SSE, HTTP and WebSocket server types, and authentication.

    38k GitHub starsUsed in 11 repos~3.1k tokens
    Agent WorkflowsAuto-check passed
  • Fastmcp Client CLI

    PrefectHQ/fastmcp

    Query and invoke tools on MCP servers using fastmcp list and fastmcp call.

    28k GitHub starsUsed in 1 repo~823 tokens
    Agent WorkflowsAuto-check passed
  • Crush Configuration

    charmbracelet/crush

    Explains how to configure the Crush coding agent with crushrc or crush.json, covering providers, models, LSPs, MCP servers, hooks, permissions and config precedence.

    29k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Context Mode Output Sandbox

    mksglu/context-mode

    Routes large command, file, API and browser output through context-mode tools so only the needed result enters the agent's context, instead of dumping it via Bash.

    26k GitHub stars~4.1k tokensUpdated today
    Agent WorkflowsAuto-check passed

More from mcpc-tech/mcpc

  • Creates presentation slides using Slidev with Markdown syntax.

    103 GitHub stars~4.6k tokensUpdated yesterday
    Auto-check passed
  • Release V5 Acp

    mcpc-tech/mcpc

    Release @mcpc/acp-ai-provider package for AI SDK v5 (0.1.x versions) from release-v5 branch.

    103 GitHub stars~657 tokensUpdated yesterday
    Auto-check passed
  • Version Upgrade

    mcpc-tech/mcpc

    Upgrades mcpc package versions and their dependencies. An agent skill from mcpc-tech/mcpc.

    103 GitHub stars~415 tokensUpdated yesterday
    Auto-check passed

Categories

Questions about Mcpc Core

What does Mcpc Core do?

Build agentic MCP servers using @mcpc/core by composing existing MCP tools. Mcpc Core is an agent skill from mcpc-tech/mcpc. Build agentic MCP servers using @mcpc/core by composing existing MCP tools.

When should I use Mcpc Core?

Mcpc Core fits situations like: debugging MCPC agentic servers — including setting up compositions; choosing execution modes; writing plugins; connecting transports.

How do I install Mcpc Core in Claude Code?

Run `npx skills add mcpc-tech/mcpc --skill mcpc-core -a claude-code`. Or copy the skill folder (.agents/skills/mcpc-core in mcpc-tech/mcpc) into .claude/skills/mcpc-core in your project. Claude Code loads it when a task matches its description.

How do I install Mcpc Core in Codex?

Run `npx skills add mcpc-tech/mcpc --skill mcpc-core -a codex`. Or copy the skill folder (.agents/skills/mcpc-core in mcpc-tech/mcpc) into .agents/skills/mcpc-core in your project. Codex loads it when a task matches its description.

Can I use Mcpc Core in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add mcpc-tech/mcpc --skill mcpc-core -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/mcpc-core, .gemini/skills/mcpc-core, .github/skills/mcpc-core and .opencode/skills/mcpc-core in your project.

What does Mcpc Core need to run?

Going by SKILL.md and its folder, Mcpc Core needs the command-line tools its instructions call (deno, npm and pnpm). Our summary lists: Node.js.

Does Mcpc Core access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Mcpc Core safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Mcpc Core use?

Mcpc Core is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Mcpc Core use?

About 2.2k tokens (SKILL.md is roughly 8.7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 1.5k tokens, read only when the agent opens those files.

What are the alternatives to Mcpc Core?

Skills that share tags, products or a category with Mcpc Core: MCP Server Builder (anthropics/skills, 180k stars), MCP Server Builder (shareAI-lab/learn-claude-code, 78k stars), MCP Integration for Plugins (anthropics/claude-plugins-official, 38k stars) and Fastmcp Client CLI (PrefectHQ/fastmcp, 28k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Mcpc Core?

mcpc-tech (a GitHub organization) maintains it in mcpc-tech/mcpc, which has 103 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on October 6, 2026.

Source: mcpc-tech/mcpc on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.