By nicobailon
Convert MCP servers into standalone CLI tools for AI agents. Auto-detects npm/PyPI packages, generates Node.js wrappers, and registers them to CLAUDE.md.
Claude Code plugin that converts MCP servers into standalone CLI tools—no context bloat, no complex syntax.
/mcp2cli chrome-devtools-mcp
Install the plugin:
/plugin install nicobailon/mcp2cli-plugin
Convert an MCP server:
/mcp2cli chrome-devtools-mcp
Use the generated tools:
chrome-snapshot --help
chrome-snapshot --selector "#main"
When you install MCP servers in Claude Code, their tool definitions are loaded into the context window before you send a single message. A server with 20+ tools can consume thousands of tokens just sitting there. This adds up quickly with multiple servers.
Even with mcporter installed, Claude Code won't proactively use these commands because:
--help for self-discoveryConvert MCP tools to simple, self-documenting CLI scripts. Claude gets the same functionality via Bash, but:
--help is called)--help| mcporter CLI | mcp2cli-generated |
|---|---|
| 100+ chars, agents fumble | ~50 chars, reliable |
No --help | Full --help with examples |
| Abstract entry point | One tool = one script |
| Requires full path | Symlinked for direct invocation |
| Not proactively used | Claude uses reliably |
# Before: Complex mcporter command Claude won't use proactively
npx mcporter call --stdio "npx -y chrome-devtools-mcp" chrome-devtools.take_snapshot
# After: Simple script Claude uses reliably
chrome-snapshot --selector "#main"
For the rationale behind this approach, see What if you don't need MCP? by Mario Zechner.
This is a prompt-only plugin—no executable code, just instructions for Claude:
/mcp2cli <package> or describe what you want naturally--help support~/.local/bin for direct invocation~/.claude/CLAUDE.md/mcp2cli chrome-devtools-mcp
/mcp2cli mcp-server-fetch
/mcp2cli @anthropic-ai/some-mcp-server
Just describe what you want:
Tools are generated to ~/agent-tools/<server-name>/:
~/agent-tools/chrome-devtools/
├── package.json
├── README.md
├── chrome-snapshot.js
├── chrome-navigate.js
└── chrome-click.js
Each script is symlinked to ~/agent-tools/bin (or ~/.local/bin as fallback) without the .js extension, so Claude can invoke by name—no full paths needed:
# Just the symlink name works
chrome-snapshot --help
chrome-navigate --url "https://example.com"
| Issue | Fix |
|---|---|
| mcporter not found | Ensure npx mcporter works |
| Discovery fails | Server may be Python—mcporter tries uvx automatically |
| Tools not in PATH | Add ~/agent-tools/bin or ~/.local/bin to your PATH |
MIT
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub nicobailon/mcp2cli-pluginGenerate beautiful HTML pages for diagrams, diff reviews, plan reviews, slides, and data tables
MCP server development helper with tool and resource scaffolding
Configurable MCP wrapper that consolidates your tools into just 3, using semantic search for on-demand discovery and sandboxed TypeScript execution. Ships with 7 example servers (96 tools) - reduces context consumption by 97% (48k tokens down to 1.1k).
Build FastMCP 3.x Python MCP servers — covers provider/transform architecture (including CodeMode, Tool Search, and server-level transforms), component versioning, session state, authorization (MultiAuth, PropelAuth, connection-pooled token verifiers), evaluation creation, Pydantic validation, async patterns, STDIO and HTTP transports, nginx reverse proxy deployment, background tasks, Prefab Apps UI, security patterns, client SDK usage, testing, deployment, and migration from FastMCP v2. TypeScript is a legacy reference only and is not updated for v3.
Manage MCP servers - discover, analyze, execute tools/prompts/resources. Use for MCP integrations, capability discovery, tool filtering, programmatic execution, or encountering context bloat, server configuration, tool execution errors.
Execute TypeScript code that calls MCP tools with full type safety
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses