{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifch5a6egzwfkusmsb5bx5soyaggsqxr432ld5fcwcn7rlslcawle",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mki254aezau2"
  },
  "path": "/t/template-for-scalable-agent-skill-creation-looking-for-feedback/1379825#post_2",
  "publishedAt": "2026-04-27T11:35:04.000Z",
  "site": "https://community.openai.com",
  "tags": [
    "GitHub - TechSpokes/skill-ide-phpstorm-mcp: IDE PHPStorm MCP server best practices. Helps AI agents to leverage your IDE tools for better efficiency. · GitHub"
  ],
  "textContent": "So I used the above repo to build a skill I needed from cloned base with a prompt:\n\n\n    # Intake research phase\n\n    I need a skill to boost AI agent efficiency while working with my IDE: PHPStorm. It has a native MCP server that is often more efficient than default agents tools available in terminal. Before building the skill (bootstrapping), I want you to test the MCP server’s functionality and compare it with terminal tools to understand the potential efficiency gains and document everything as the input for the skill development.\n\n    Tasks\n\n    IMPORTANT: do not start the skill build phase until you have completed the testing and documentation of the MCP server’s functionality and efficiency compared to terminal tools!\n\n    Could you install the MCP server from the .intake/.mcp.json file and then create a simple sample php app (inside .intake/app/ folder - choose the app purpose at your liking) for testing the MCP server’s functionality and compare the agent experience and efficiency between using the MCP server and terminal tools you have?\n\n    NOTE: Read the MCP server tools to understand what is available to help you craft the sample app features and functionalities as well as the scenarios for testing the tools. Goal is to cover as many as possible to get the complete picture of the MCP server’s capabilities and efficiency compared to terminal tools.\n\n    Documentation and Reporting\n\n    While working on the app document your experience in the .intake/experience/ folder, try to group themes of functionalities, efficiency, and any issues you encounter. Once you have the app working, create a report comparing the agent experience and efficiency between using the MCP server and terminal tools, and save it in the .intake/experience/ folder as well.\n\n    Planning\n\n    If needed plan your work in the .intake/mcp-usage-testing.plan.yaml and .intake/php-app-development.plan.yaml files to outline your approach for both tasks.\n\n\nThe new skill is here (check history to see the whole process): GitHub - TechSpokes/skill-ide-phpstorm-mcp: IDE PHPStorm MCP server best practices. Helps AI agents to leverage your IDE tools for better efficiency. · GitHub\n\nBasically the intake research (I did not have any files to show to the skill builder took about an hour), but skill building and packaging took about 10 min… and the update took another 5 min of my time.\n\nHave fun!",
  "title": "Template for Scalable Agent Skill Creation (Looking for Feedback)"
}