Skip to content

AI coding execution control skills | AI 编程执行约束技能

License

Notifications You must be signed in to change notification settings

trae-community/trae-skills

TRAE Agent Skills

License: MIT PRs Welcome

Community-maintained Agent Skills for TRAE. In TRAE, a skill is a reusable, scenario-specific “capability manual” defined by a SKILL.md file, optionally packaged with scripts, templates, examples, and other resources. The agent scans skill descriptions first and only loads full skill content when a task is highly relevant, reducing token usage and avoiding irrelevant context.

中文说明

Quickstart

  1. Clone this repository.
  2. Put skills into the right location for TRAE:
    • Project skills: .trae/skills/<skill-name>/SKILL.md
    • Global skills: ~/.trae/skills/<skill-name>/SKILL.md
  3. Refresh TRAE's skill discovery in settings (location varies by TRAE version).
  4. Ask TRAE for a task that matches a skill’s description, for example:
    • “Use the webapp-testing skill to create Playwright tests for the login flow.”
    • “Use the release-notes skill to draft release notes from recent PR titles.”

What are Agent Skills?

Agent Skills are folders of instructions, scripts, and resources that an AI agent can discover and load dynamically to perform specialized tasks in a repeatable way. A typical skill is a directory with a SKILL.md file that contains:

  • YAML frontmatter metadata (especially name and description)
  • A Markdown body with steps, guidelines, and examples

This pattern keeps the agent’s core rules lightweight while making SOP-style workflows portable and shareable.

Skills vs. other features in TRAE

  • Skills vs. Rules: rules are fully injected into every chat and continuously occupy context; skills are loaded on-demand only when called.
  • Skills vs. MCP servers: skills describe how TRAE should accomplish a task; MCP servers provide tools that TRAE can call. For example, a Playwright MCP server provides browser automation tools, while a testing skill defines test structure, conventions, and execution workflow.

Skill types in TRAE

  • Global skills: reusable across projects (personal/team conventions, general toolchain workflows, long-term output preferences).
  • Project skills: apply only to the current project (project-specific business rules, architecture constraints, project scaffolding/testing workflows).

Repository layout

This repository is intended to follow a simple, discoverable layout:

skills/
  _template/               # Template for creating new skills
    SKILL.md
  <skill-name>/
    SKILL.md               # (Mandatory) Core instructions for the agent
    (optional) examples/   # Input/output examples
      input.md
      output.md
    (optional) templates/  # Reusable templates
      component.tsx
    (optional) resources/  # Reference files, scripts, or assets
      style-guide.md

Skill format (SKILL.md)

Each skill must include a SKILL.md with YAML frontmatter:

---
name: skill-name
description: Brief description of the skill's function and usage scenario.
---

# Skill Name

## Description
Describe what this skill does.

## Usage Scenario
Describe the conditions that trigger this skill.

## Instructions
Clear, step-by-step instructions telling the agent exactly what to do.

## Examples (Optional)
Input/output examples showing the expected result.

Guidelines for good metadata:

  • name: lowercase, use hyphens instead of spaces, and keep it stable over time
  • description: be specific about what it does and when to use it (this is what helps agents decide to load the skill)

Skills catalog

This section will list available skills as they are added.

Skill Description Usage Scenario Status
git-commit-generator Generate standardized git commit messages based on code changes (diffs), following Conventional Commits specification. Git Operations, Code Review Stable

Tip: To add your skill to this catalog, update this table in your PR.

Contributing

See CONTRIBUTING.md.

License

See LICENSE.

Disclaimer

Skills in this repository are provided for community/educational use. Always review and test skills in your own environment before relying on them for production or security-sensitive workflows.

Links

About

AI coding execution control skills | AI 编程执行约束技能

Resources

License

Code of conduct

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors