Agent-Skills.md

Agent Skills: MCP Management

[Tooling & Meta] Manage Model Context Protocol (MCP) servers - discover, analyze, and execute tools/prompts/resources from configured MCP servers. Use when working with MCP integrations, need to discover available MCP capabilities, filter MCP tools for specific tasks, execute MCP tools programmatically, access MCP prompts/resources, or implement MCP client functionality. Supports intelligent tool selection, multi-server management, and context-efficient capability discovery.

UncategorizedID: duc01226/easyplatform/mcp-management

Repository

duc01226/easyplatform
duc01226License: MIT
31

Install this agent skill to your local

pnpm dlx add-skill https://github.com/duc01226/EasyPlatform/tree/HEAD/.claude/skills/mcp-management

Skill Files

Browse the full folder contents for mcp-management.

Download Skill

Loading file tree…

.claude/skills/mcp-management/SKILL.md

Skill Metadata

Name
mcp-management
Description
"[Tooling & Meta] Manage Model Context Protocol (MCP) servers - discover, analyze, and execute tools/prompts/resources from configured MCP servers. Use when working with MCP integrations, need to discover available MCP capabilities, filter MCP tools for specific tasks, execute MCP tools programmatically, access MCP prompts/resources, or implement MCP client functionality. Supports intelligent tool selection, multi-server management, and context-efficient capability discovery."

MCP Management

Manage and interact with Model Context Protocol (MCP) servers.

Overview

MCP enables AI agents to connect to external tools and data sources. This skill provides discovery, analysis, and execution of MCP capabilities.

Key Benefits: Progressive disclosure, intelligent tool selection, multi-server management, persistent tool catalog (assets/tools.json).

When to Use

  1. Discovering MCP capabilities (tools/prompts/resources)
  2. Task-based tool selection
  3. Executing MCP tools programmatically
  4. Building/debugging MCP client implementations
  5. Context management (delegate MCP ops to subagents)

Configuration

MCP servers configured in .claude/.mcp.json.

Gemini CLI Integration: mkdir -p .gemini && ln -sf .claude/.mcp.json .gemini/settings.json

GEMINI.md: Auto-loaded by Gemini CLI, enforces structured JSON responses:

{"server":"name","tool":"name","success":true,"result":<data>,"error":null}

See references/configuration.md and references/gemini-cli-integration.md.

Execution Priority

1. Gemini CLI (Primary)

# IMPORTANT: Use stdin piping, NOT -p flag (deprecated, skips MCP init)
echo "Take a screenshot of https://example.com. Return JSON only per GEMINI.md instructions." | gemini -y -m gemini-2.5-flash

Check availability: command -v gemini

2. Direct CLI Scripts (Secondary)

cd .claude/skills/mcp-management/scripts && npm install
npx tsx cli.ts list-tools       # Saves to assets/tools.json
npx tsx cli.ts list-prompts
npx tsx cli.ts list-resources
npx tsx cli.ts call-tool <server> <tool> <json>

3. mcp-manager Subagent (Fallback)

Use when Gemini CLI unavailable. Keeps main context clean.

Implementation Patterns

| Pattern | When | How | | -------------------------- | ------------------------- | ---------------------------------------------- | | Gemini CLI Auto-Execution | Default (fastest) | echo "task" \| gemini -y -m gemini-2.5-flash | | LLM-Driven Tool Selection | Need intelligent matching | LLM reads assets/tools.json | | Multi-Server Orchestration | Cross-server coordination | Tools tagged with source server | | Subagent Delegation | Context efficiency | mcp-manager agent handles MCP ops |

Scripts Reference

| Script | Purpose | | ----------------------- | --------------------------------------------------------------------------- | | scripts/mcp-client.ts | Core MCP client (config, connect, list, execute) | | scripts/cli.ts | CLI interface (list-tools, list-prompts, list-resources, call-tool) |

Technical Details

See references/mcp-protocol.md for JSON-RPC protocol, message types, error codes, transports.

IMPORTANT Task Planning Notes

  • Always plan and break many small todo tasks
  • Always add a final review todo task to review the works done at the end to find any fix or enhancement needed