DEV Community: AI-Hub-Admin

Tired of All the Agent Access Formats And I Built a OneKey Gateway as omni converter to Ship APIs as MCPs Skills CLIs 10x faster

AI-Hub-Admin — Thu, 09 Apr 2026 04:55:31 +0000

Hey dev community 👋

As the agent ecosystem keeps growing, I’ve noticed a recurring problem:

You might have one solid business API or proprietary data, but to make it usable across different agent platforms, you end up rebuilding it in multiple formats:
MCPs, Skills ,CLIs , REST APIs , Even language-specific SDKs (Python / TypeScript). That is painful when you want to upgrade, you have to upgrade on all of it, which means duplicated effort, multiple codebases, and painful maintenance.

That’s why I started OneKey Gateway project:
GitHub OneKey Gateway

with internal gateway layer, central Agent API registry platform, onekey authentication, and omni format converter, you only need to maintain one API/data, and users/agents can use in all formats.

Introduction

OneKey Gateway Agent/Skills/MCP Conversion Matrix Roadmap

Register and maintain your APIs only once, allow agents/users to use your APIs in other formats 10x faster, supported CLIs/Skills/Rest API/MCPs

From \ To	API	CLI	Skills	Routed API	MCPs (StreamingHTTP)	MCPs (Stdio Local)
API	-	✅	✅	✅	✅	-
CLI	-	-	✅	-	-	✅
Skills	-	-	-	-	✅	✅
Routed API	-	-	-	-	-	-
MCPs (Streaming HTTP)	-	-	-	-	-	-
MCPs (Stdio Local)	-	-	-	-	-	-

So the question is: Today, most solutions only partially solve this. You’ll find tools like fastapi_mcp and similar converters—but they’re still code-level solutions:
You still need to write glue code. You still need to maintain multiple packages, publish and version each format
In reality, you might end up managing 6–10 different packages for a single API.
That’s not scalable.

The workflow of OneKey Agent Gateway works like

Register your API 
→ Store metadata (Agent meta, API meta OpenAPI, auth, configs)
→ Auto-generate & distribute as
   - CLI
   - Skills
   - MCP
   - REST Router
-> Agent Requests Routed to your APIs

Here is the roadmap of potential cases of conversion between various formats: https://github.com/aiagenta2z/onekey-gateway
Now I will introduce you how to convert a food calories API to various formats without publishing any more packages and use the OneKey Gateway.

OneKey Gateway Registered and Routed

Original API

Food Nutrition API from U.S. Department of Agriculture

Search the nutritions caloeries of Cheddar Cheese.

curl -XPOST -H "Content-Type:application/json" -d "{\"query\":\"Cheddar cheese\"}" https://api.nal.usda.gov/fdc/v1/foods/search?api_key=DEMO_KEY

After registration as in docs OneKey Gateway Doc. Agents/Users can use your APIs in various formats:

Convert API to CLI

npx onekey agent fdcnal/usda-fooddata-central-agent search_foods '{"query": "Cheddar Cheese","pageSize": 10}'

Convert API to Skills

npx agtm skills build fdcnal/usda-fooddata-central-agent

It will generate local ./skills folder with scripts and CLIs

Convert API to MCP

npx onekey mcp fdcnal/usda-fooddata-central-agent

It generates a mcp config to access your APIs

{
  "mcpServers": {
    "deepnlp-onekey-fdcnal-usda-fooddata-central-agent": {
      "url": "https://agent.deepnlp.org/mcp?server_name=fdcnal/usda-fooddata-central-agent&onekey=YOUR_KEY"
    }
  }
}

Convert API to REST Router in programming way

export DEEPNLP_ONEKEY_ROUTER_ACCESS=YOUR_ROUTER_KEY
curl -X POST "https://agent.deepnlp.org/agent_router" -H "Content-Type: application/json" -H "X-OneKey: $DEEPNLP_ONEKEY_ROUTER_ACCESS" -d '{"unique_id":"fdcnal/usda-fooddata-central-agent","api_id":"search_foods","data":{"query":"Cheddar Cheese","pageSize":10}}'

And for the remaining blog, I will introduce you how to register the API to DeepNLP x AI Agent A2Z agent registry.
and distribute in various formats.

2. Quickstart Example: USDA FoodData Central API

Let’s see how to register and convert a Food Nutrition API:

2.1 Prepare API Specs

Use the USDA OpenAPI specs (fdc_api.json or fdc_api.yaml) to define your agent metadata. If no OpenAPI or similar specs are available, provide the python/typescript definition as "def my_func(param1: str,...)" to coding agent/claude code/codex to help you generate the meta.

agent.yaml

api_list:
  - api_id: search_foods
    protocol: https
    description: Search for food items and calories
    endpoint: https://api.nal.usda.gov/fdc/v1/foods/search
    method: POST
    params:
      query: str
      pageSize: int
    auth:
      type: API_KEY
      header: "X-Api-Key"
      value: $YOUR_FDC_API_KEY

You can also generate this using Claude Code or Codex. I am using the prompt, in the Github repo run the prompt:

Generate agent.json and agent.yaml file following the format of ./examples/usda_food_api/agent.json and agent.yaml format,
fill in the api_list key, with meta of APIs: api_id,protocol,description,endpoint,method,params,auth, For params, you can read the necessary informations from the the OpenAPI specs of fdc_api.json.
Output to /agent.json or /agent.yaml formats

1.2 Register Your API

Get your OneKey Gateway registry key, the key will identify you as the owner of the registered API. Get Keys at Keys

export AI_AGENT_MARKETPLACE_ACCESS_KEY=YOUR_REGISTRY_KEY
npx agtm upload --config ./agent.yaml

Your agent’s unique ID is in user_id/repo_name format, e.g.:

fdcnal/usda-fooddata-central-agent

1.3. Usage Examples

API2REST

The request will be posted to central registry endpoint, https://agent.deepnlp.org/agent_router, authenticated once and routed
to your registered API endpoint with payload and headers for authentication. Note the access_key devs provide are encrypted safely and decrypted/routed to your registered API.

export AI_AGENT_MARKETPLACE_ACCESS_KEY=YOUR_REGISTRY_KEY

curl -X POST "https://agent.deepnlp.org/agent_router" \
-H "Content-Type: application/json" \
-H "X-OneKey: $DEEPNLP_ONEKEY_ROUTER_ACCESS" \
-d '{
  "unique_id": "fdcnal/usda-fooddata-central-agent",
  "api_id": "search_foods",
  "data": {"query": "Cheddar Cheese","pageSize": 10}
}'

API2CLI

Users can use your API in the common command line npx onekey agent <unique_id> <api_id> <payload>

npx onekey agent <unique_id> <api_id> <payload>
npx onekey agent fdcnal/usda-fooddata-central-agent search_foods '{"query": "Cheddar Cheese","pageSize": 10}'

API2Skills

If you readily convert your APIs to local skills under the folder ./skills/your-api-converted-skill/*

npx agtm skills build fdcnal/usda-fooddata-central-agent

Generates:

skills/fdcnal-usda-fooddata-central-agent/
  ├── SKILL.md
  ├── scripts/
  └── reference/

API2MCP Server

npx onekey mcp fdcnal/usda-fooddata-central-agent

Generates config for StreamingHTTP MCPs:

{
  "mcpServers": {
    "deepnlp-onekey-fdcnal-usda-fooddata-central-agent": {
      "url": "https://agent.deepnlp.org/mcp?server_name=fdcnal/usda-fooddata-central-agent&onekey=YOUR_KEY"
    }
  }
}

RoadMap

If you like the ideas of the onekey gateway and roadmap of converter.
You can also contribute more use cases, documents and your registered APIs at the Onekey Gateway GitHub and AI Agent Marketplace GitHub.

I Keep Forgetting CLIs So Does My AI Agent - Simplifying CLI Usage for Human and AI Agents with Hints and Completion

AI-Hub-Admin — Fri, 20 Mar 2026 08:07:52 +0000

As an LLM/AI Agent application developer, I frequently use tools like Claude Code, Codex, or OpenClaw to manage agent tasks via MCP, Skills, and especially command-line interfaces (CLIs).

One problem I face: I can never remember all the exact CLI commands, positional arguments, or options for tools like playwright-cli, claude, or codex. Searching through documentation or README.md files every time is tedious
and I am just too lazy to do it.

To solve this, I spend some time to built a CLI helper tool with hints and auto-completion capabilities.
This tool helps both humans and agents quickly discover and run CLI commands without memorizing them. And I would like to share with the community
agtm which means agent manager or agent marketplace.

How It Works

For example, if you want to test playwright-cli for web page reading but forget the exact commands:

Type a keyword, e.g., play, after agtm run:

npx agtm run play

Then you can follow the guidance and cli hints to finalize your cli.

Step 1. Choose the suggested Skill ID

Skill ID suggestions:
  1. microsoft/playwright-cli
  2. paul-gauthier/aider
  3. anthropic/claude-code
  4. openai/codex-cli
  5. googleworkspace/cli

Step 2. Choose the suggested CLI:

Command hints:
  1. playwright-cli goto <url> # navigate to a url
  2. playwright-cli open [url] # open browser, optionally navigate to url

Step 3.
Complete the final command by filling in parameters:

playwright-cli goto https://www.google.com
agtm run microsoft/playwright-cli playwright-cli goto https://www.google.com

Then it gives you exact cli running results

### Ran Playwright code

js
await page.goto('https://www.google.com');

### Page
- Page URL: https://www.google.com/
- Page Title: Google
### Snapshot
- [Snapshot](../aiagenta2z/agtm/.playwright-cli/page-2026-03-20T08-03-59-159Z.yml)

shell

Installation

npm install -g @aiagenta2z/agtm
npx agtm setup --hint  # Builds a local RAG of CLI hints
npx agtm run <provider_unique_id> <agent_cli>

provider_unique_id: identifies the project owner/repo (GitHub style)
agent_cli: the CLI commands in that repository

Agent Mode

Agents like Claude or Codex can also use the hint system to run commands automatically if there are documents mismatch and skills update. In agent mode, the tool searches relevant Skill IDs and CLI suggestions for non-interactive execution:

npx agtm run play --mode agent

Skill ID suggestions:
  1. microsoft/playwright-cli
     - playwright-cli open [url]  # open browser, optionally navigate to url
     - playwright-cli goto <url>   # navigate to a url

prompt: Help me visit producthunt.com and extract front-page items to a local text file.

Architecture

CLI RAG Local:
agtm setup builds a local RAG (trie) of CLI hints. It matches Skill IDs first, then provides CLI suggestions and hints for the selected Skill.

Agent Mode:
Enables agents to use CLI hints automatically. Instead of interactive selection, the tool provides relevant CLI suggestions from the local RAG.

What do you think of the cli hint tool that work for both human and agent?

Please welcome to contribute to github aiagenta2z/agtm of the clis repos.