How to Use Skills in VSCode's Codex Extension

Zenn tutorial on configuring and using Skills in VSCode's Codex extension to help coding assistants better understand project context and coding standards.

Skills are like CLAUDE.md — project-specific instructions and constraints for AI coding assistants. Article shows configuration methods and common Skill templates.

Significantly improves AI-assisted efficiency for engineers using Codex daily.

The core concept of Skills is providing AI Coding assistants with project-specific context so generated code directly follows project conventions. The author's comparative tests show a 50-70% reduction in manual code modifications after configuring Skills. This approach of 'teaching AI to understand your project' is key to making Vibe Coding work in real-world projects. Essential reading for developers using AI coding assistants daily.

Codex is OpenAI's VSCode AI coding assistant extension. Skills make it better understand your project.

What Are Skills

Skills are predefined instruction files telling Codex about your project: frameworks used, coding standards, directory structure, test requirements. Similar to CLAUDE.md in the Anthropic ecosystem.

Configuration

Create `.codex/skills/` directory at project root, each Skill is a markdown file:

.codex/
skills/
project.md     # Project overview and architecture
coding.md      # Coding standards
testing.md     # Test requirements
deployment.md  # Deployment process

Common Templates

project.md: Describe tech stack, directory structure, core modules, dependencies.

coding.md: Define code style (indentation, naming, comments), error handling patterns, log formats.

testing.md: Specify test framework, coverage requirements, mock strategies, test file organization.

Before/After Comparison

  • Without Skills: Codex generates code that doesn't match project conventions, needs manual adjustment
  • With Skills: Generated code directly matches project standards, reducing manual edits by 50-70%

Best Practices

  • Keep Skills files concise (<500 lines total)
  • Include specific code examples (not just text descriptions)
  • Update regularly to reflect project evolution
  • Share unified Skills files across the team

Industry Trend Connection

Project-level context configuration is becoming standard practice in AI Coding—Anthropic's CLAUDE.md, OpenAI's Codex Skills, and Cursor's rules all serve the same purpose. This is a crucial step for Vibe Coding moving from 'toy' to 'productivity tool'. MCP protocol is also evolving in this direction, aiming to provide standardized project context interfaces for AI Agents.

In-Depth Analysis and Industry Outlook

From a broader perspective, this development reflects the accelerating trend of AI technology transitioning from laboratories to industrial applications. Industry analysts widely agree that 2026 will be a pivotal year for AI commercialization. On the technical front, large model inference efficiency continues to improve while deployment costs decline, enabling more SMEs to access advanced AI capabilities. On the market front, enterprise expectations for AI investment returns are shifting from long-term strategic value to short-term quantifiable gains.

However, the rapid proliferation of AI also brings new challenges: increasing complexity of data privacy protection, growing demands for AI decision transparency, and difficulties in cross-border AI governance coordination. Regulatory authorities across multiple countries are closely monitoring these developments, attempting to balance innovation promotion with risk prevention. For investors, identifying AI companies with truly sustainable competitive advantages has become increasingly critical as the market transitions from hype to value validation.