HN Showcase
Back to browse

I built an AI Agent skill translation and refactoring tool

by nicholas_pw·Mar 3, 2026·1 point·1 comment
View on HN

AI Analysis

●MidSolve My Problem

Saves hours of JSON schema rewriting, but the agent ecosystem is still too young to tell if this sticks.

Strengths
  • •Directly addresses real friction point—manually rewriting agent tool schemas across frameworks is tedious and error-prone.
  • •Automates data-type mapping logic that's currently done by hand or ad-hoc scripts.
Weaknesses
  • •Early-stage, community-dependent tool; unclear which frameworks are actually supported and how complete the mappings are.
  • •Agent frameworks are nascent and still consolidating; this tool's value is hostage to ecosystem maturation and API stability.
Target Audience

AI agent framework users, solo builders working across multiple agent ecosystems

Similar To

Anthropic's tool-use spec · OpenAI's function calling docs

Post Description

Hey HN,

I’m a solo builder and I've been doing a lot of work with AI Agents lately. I realized that as the ecosystem grows, moving skills (or tools) between different agent frameworks is a massive headache. You often have to rewrite the JSON/JSONB schemas or adapt the logic completely.

To solve my own scratching, I built skills-refiner to translate and refine AI Agent Skill

It’s a tool designed to automatically translate and refactor Agent skills. You feed it a skill definition from one framework, and it spits out the correctly formatted schema for another, handling the underlying data type mappings.

It's still in the early stages, but it's already saving me hours of manual JSON editing.

I'd love to get feedback from the community. What agent frameworks are you using most right now? And what's your biggest pain point when dealing with tool calling schemas?

Similar Projects