Twig
Score and tune your own MCP servers before listing them so agents get clearer tools and you ship a higher-quality integration.
Overview
Twig is a MCP server for the Build phase that scores and optimizes MCP server quality for agent-facing tool catalogs.
What is this MCP server?
- MCP server quality scoring tuned for the agent economy
- Optimization guidance to improve tool discoverability and reliability
- stdio npm package @kind-ling/twig (v0.4.2) for local agent wiring
- GitHub source at Kind-ling/twig for iteration alongside your server repo
- Registry version 0.4.2
- npm identifier @kind-ling/twig with stdio transport
- Source repository github.com/Kind-ling/twig
What problem does it solve?
You shipped an MCP server but agents misfire on tools, descriptions feel thin, and you have no structured way to know if it is marketplace-ready.
Who is it for?
Indie builders publishing or maintaining MCP servers who want a feedback loop before public registry or directory listing.
Skip if: Teams with no MCP server to evaluate, or builders who only consume third-party MCP and never author their own tools.
What do I get? / Deliverables
After wiring Twig, you get actionable quality signals and optimization steps so your MCP server is clearer, more reliable, and easier for agents to invoke correctly.
- Quality scores and optimization notes for your MCP server
- Clearer improvement targets for tool names, schemas, and descriptions
- Higher confidence before publishing to registries or directories
Recommended MCP Servers
Journey fit
Twig targets builders who publish MCP servers as part of product or internal agent stacks, which is core agent-tooling work in the Build phase. Quality scoring and optimization sit next to skill and MCP authoring—improving tool contracts, metadata, and behavior before agents depend on them.
How it compares
MCP quality auditor for server authors, not a general coding agent skill or runtime integration for end-user apps.
Common Questions / FAQ
Who is Twig for?
Twig is for developers who build and maintain MCP servers and want structured quality scoring before agents or catalogs rely on their tools.
When should I use Twig?
Use Twig while iterating a new MCP server, before registry submission, or after refactors when tool descriptions and behavior need a consistency check.
How do I add Twig to my agent?
Install the npm package @kind-ling/twig, add it as a stdio MCP server in Claude Code, Cursor, or Codex, then invoke its scoring tools against your target server project.