Plugin

claudikins-tool-executor

Name: claudikins-tool-executor
Author: elb-pr

Consolidate 96+ MCP tools into a 3-tool workflow for Claude Code: semantically search tools across configurable servers, fetch schemas, and execute them in a sandboxed TypeScript environment. Configure servers, diagnose issues with tests and health checks, and follow guided usage patterns to slash context tokens by 97%.

Component Overview

/te-config, /te-doctor +1

Commands

tool-executor-guide

Agents

te-config, te-doctor +2

Skills

SessionStart, UserPromptSubmit

Hooks

tool-executor

MCP Servers

LSP Servers

Output Styles

Install

npx claudepluginhub elb-pr/claudikins-marketplace --plugin claudikins-tool-executor

Component Details

Commands (3)

claudikins-tool-executor:config

/te-config

Configure Tool Executor - add/remove MCP servers, set environment variables

claudikins-tool-executor:doctor

/te-doctor

Diagnose Tool Executor issues - run tests, check health, troubleshoot

claudikins-tool-executor:guide

/te-guide

Learn how to use the Tool Executor - 3-tool workflow, workspace API, MCP clients

Agents (1)

tool-executor-guide

/tool-executor-guide

Use this agent when users ask about the Claudikins Tool Executor MCP server - how to use it, configure it, or troubleshoot it. This agent has deep knowledge of the codebase and can help with usage patterns, adding MCP servers, and diagnosing issues. <example> Context: User is trying to use execute_code but getting errors user: "execute_code keeps timing out, what's wrong?" assistant: "I'll use the tool-executor-guide agent to diagnose the timeout issue." </example> <example> Context: User wants to add a new MCP server to the sandbox user: "How do I add a new MCP server to tool executor?" assistant: "Let me use the tool-executor-guide agent to walk you through adding a new MCP server." </example> <example> Context: User is confused about the workspace API user: "What methods are available on workspace?" assistant: "I'll use the tool-executor-guide agent to explain the workspace API." </example> <example> Context: User wants to understand how search_tools works user: "How does search_tools find the right tools?" assistant: "Let me use the tool-executor-guide agent to explain the search architecture." </example>

Skills (4)

te-config

/skills/te-config

Use this skill when users ask "add MCP server", "remove MCP", "configure tool executor", "add new tool", "environment variables for tool executor", "regenerate registry", or need to modify which MCP servers are available in the sandbox.

te-doctor

/skills/te-doctor

Use this skill when users say "tool executor not working", "diagnose tool executor", "check MCP health", "run tool executor tests", "debug search_tools", "execute_code failing", or need to troubleshoot issues with the Tool Executor.

te-guide

/skills/te-guide

Use this skill when users ask "how does tool executor work", "how to use execute_code", "workspace API", "MCP client examples", "search_tools examples", "context-efficient patterns", or need guidance writing code for the Tool Executor sandbox.

using-tool-executor

/skills/using-tool-executor

Use when a task needs capabilities BEYOND basic tools (Read/Grep/Glob/Bash) - semantic search, AI analysis, research, image generation, etc.

Hooks (1)

Review workflow modifications before installing

Event Hooks

2 hooks across 2 events

MCP Servers (1)

Connects to external services

tool-executor

admin

README

Tool Executor

Programmatic MCP Execution for Claude Code
The API has batched tool calling. Claude Code gets serial execution. This bridges the gap.

Quick Start • The 3-Tool Workflow • Wrapped Servers • Configuration • Roadmap

Anthropic's API users get programmatic tool calling - Claude writes code, executes N tools in a sandbox, returns once. Claude Code users get serial execution and lazy loading. Tool Executor brings the API pattern to Claude Code.

Aspect	Claude Code (stable)	Claude Code 2.1.7	Tool Executor
Schema Loading	All upfront	Lazy (>10% threshold)	Lazy (search on demand)
Execution	Serial (pause per tool)	Serial (pause per tool)	Batched (N tools, 1 return)
Output Handling	Dumps to context	Dumps to context	Auto-saves to workspace
Tool Awareness	All schemas visible	"Search available"	Hook-injected guidance

Context savings: ~97% reduction (48k to 1.1k tokens) for multi-tool workflows.

Quick Start

# Add the Claudikins marketplace
/marketplace add elb-pr/claudikins-marketplace

# Install the plugin
/plugin install claudikins-tool-executor

Restart Claude Code. Done.

First Workflow

Search for image generation tools, then generate a robot writing documentation.

Claude will:

Use search_tools to find relevant tools
Use get_tool_schema to load the exact parameters
Use execute_code to run the generation in one shot

The 3-Tool Workflow

Tool Executor exposes exactly 3 tools. Everything else happens inside the sandbox.

flowchart LR
    A[search_tools] --> B[get_tool_schema] --> C[execute_code]
    C --> D{Result > 200 chars?}
    D -->|Yes| E[Auto-save to workspace]
    D -->|No| F[Return inline]

`search_tools` - Find by Intent

Semantic search across 96 wrapped tools. Serena powers the search with BM25 fallback.

{ "query": "generate images", "limit": 5 }

Returns slim results: name, server, 80-char description. No schemas loaded until needed.

`get_tool_schema` - Load on Demand

Fetch the full JSON Schema for a specific tool before calling it.

{ "name": "gemini_generateContent" }

Returns the complete inputSchema plus usage examples.

`execute_code` - Run in Sandbox

TypeScript execution with pre-connected MCP clients. Write code that calls multiple tools, loops, branches - returns once.

const result = await gemini["gemini_generateContent"]({
  prompt: "A robot writing documentation",
  aspectRatio: "16:9"
});

// Large responses auto-save to workspace
if (result._savedTo) {
  const full = await workspace.readJSON(result._savedTo);
  console.log("Generated:", full);
}

How It Works

View full README on GitHub

Similar Plugins

everything-claude-code

Complete collection of battle-tested Claude Code configs from an Anthropic hackathon winner - agents, skills, hooks, rules, and legacy command shims evolved over 10+ months of intensive daily use

Stats

Version1.0.1

Stars44

MaintenanceExcellent

LicenseMIT

AddedJan 17, 2026

Actions

View on GitHub View README Plugin Marketplace JSON

Available In

claudikins-marketplace13

Safety Signals

Critical

Admin access level

Server config contains admin-level keywords

Caution

Uses power tools

Uses Bash, Write, or Edit tools

Tool Executor

Programmatic MCP Execution for Claude Code
The API has batched tool calling. Claude Code gets serial execution. This bridges the gap.

Quick Start • The 3-Tool Workflow • Wrapped Servers • Configuration • Roadmap

Aspect	Claude Code (stable)	Claude Code 2.1.7	Tool Executor
Schema Loading	All upfront	Lazy (>10% threshold)	Lazy (search on demand)
Execution	Serial (pause per tool)	Serial (pause per tool)	Batched (N tools, 1 return)
Output Handling	Dumps to context	Dumps to context	Auto-saves to workspace
Tool Awareness	All schemas visible	"Search available"	Hook-injected guidance

Context savings: ~97% reduction (48k to 1.1k tokens) for multi-tool workflows.

Quick Start

# Add the Claudikins marketplace
/marketplace add elb-pr/claudikins-marketplace

# Install the plugin
/plugin install claudikins-tool-executor

Restart Claude Code. Done.

First Workflow

Search for image generation tools, then generate a robot writing documentation.

Claude will:

Use search_tools to find relevant tools
Use get_tool_schema to load the exact parameters
Use execute_code to run the generation in one shot

The 3-Tool Workflow

Tool Executor exposes exactly 3 tools. Everything else happens inside the sandbox.

flowchart LR
    A[search_tools] --> B[get_tool_schema] --> C[execute_code]
    C --> D{Result > 200 chars?}
    D -->|Yes| E[Auto-save to workspace]
    D -->|No| F[Return inline]

`search_tools` - Find by Intent

Semantic search across 96 wrapped tools. Serena powers the search with BM25 fallback.

{ "query": "generate images", "limit": 5 }

Returns slim results: name, server, 80-char description. No schemas loaded until needed.

`get_tool_schema` - Load on Demand

Fetch the full JSON Schema for a specific tool before calling it.

{ "name": "gemini_generateContent" }

Returns the complete inputSchema plus usage examples.

`execute_code` - Run in Sandbox

TypeScript execution with pre-connected MCP clients. Write code that calls multiple tools, loops, branches - returns once.

const result = await gemini["gemini_generateContent"]({
  prompt: "A robot writing documentation",
  aspectRatio: "16:9"
});

// Large responses auto-save to workspace
if (result._savedTo) {
  const full = await workspace.readJSON(result._savedTo);
  console.log("Generated:", full);
}

claudikins-tool-executor

Component Overview

Install

Component Details

Commands (3)

Agents (1)

Skills (4)

Hooks (1)

MCP Servers (1)

README

Tool Executor

Quick Start

First Workflow

The 3-Tool Workflow

search_tools - Find by Intent

get_tool_schema - Load on Demand

execute_code - Run in Sandbox

How It Works

Similar Plugins

everything-claude-code

claudikins-tool-executor

Component Overview

Install

Component Details

Commands (3)

Agents (1)

Skills (4)

Hooks (1)

MCP Servers (1)

README

Tool Executor

Quick Start

First Workflow

The 3-Tool Workflow

search_tools - Find by Intent

get_tool_schema - Load on Demand

execute_code - Run in Sandbox

How It Works

Similar Plugins

everything-claude-code

team-skills-platform

superpowers

planning-with-files

performance-optimizer

conductor

`search_tools` - Find by Intent

`get_tool_schema` - Load on Demand

`execute_code` - Run in Sandbox

`search_tools` - Find by Intent

`get_tool_schema` - Load on Demand

`execute_code` - Run in Sandbox