MCP Schema Trimmer

Overview

mcp-schema-trimmer helps OpenClaw builders spot oversized MCP and tool schemas that waste context window and inflate token costs. It measures schema size, highlights the biggest noisy sections, and suggests what to trim, condense, or gate behind optional flags without breaking the core contract.

Key Features

• Measures MCP/tool schema size
• Highlights the largest noisy schema sections
• Suggests trim, condense, or optional-gating actions
• Designed to reduce token waste without breaking core contracts

Use Cases

• MCP server builders reducing token bloat
• Teams debugging noisy or expensive tool schemas
• Developers reviewing large JSON schemas before production

Requirements

• Python 3
• Valid JSON schema or tool payload file for analysis
• No external Python dependencies, standard library only
• OpenClaw v2026.3.23 or later recommended
• No API keys required

Example Usage

python3 scripts/mcp_schema_trimmer.py --schema path/to/schema.json --top 5

Getting Started

After purchase, download and extract the skill package, then copy it into your OpenClaw skills directory.

cp -r mcp-schema-trimmer/ ~/.openclaw/workspace/skills/

Documentation

Full usage notes are included with the skill package. Check the bundled SKILL.md and scripts for exact commands and configuration.

Support

Have questions? Visit GetAgentIQ.ai for documentation and support channels.