菜单
Use when writing a git commit message. Proposes a clear, Conventional Commits-compliant message from the staged changes. Invoke with: /conventional-commit, commit message, write commit, git commit, stage and commit.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | conventional-commit |
| description | Use when writing a git commit message. Proposes a clear, Conventional Commits-compliant message from the staged changes. Invoke with: /conventional-commit, commit message, write commit, git commit, stage and commit. |
| argument-hint | Optional: paste the output of 'git diff --staged' if not already in context |
| user-invocable | true |
Propose a Conventional Commits-compliant commit message from staged changes.
This workflow is packaged as a skill rather than a plain prompt because it
depends on reusable reference material in
references/commit-examples.md. Keeping the
examples outside the main file makes the guidance easier to maintain and keeps
the output more stable across repeated invocations.
git diff --staged to read the current staged changes.Determine the primary change type from the diff:
| Type | When to use |
|---|---|
feat | New feature or capability added |
fix | Bug fix |
docs | Documentation only |
style | Formatting, whitespace — no logic change |
refactor | Code restructure without behaviour change |
test | Adding or fixing tests |
chore | Build system, tooling, configuration |
Refer to references/commit-examples.md for examples.
feat(train): ...Format:
<type>(<scope>): <short description in imperative mood>
[optional body: why this change was made]
[optional footer: BREAKING CHANGE: or Closes #issue]
Rules:
2–3 candidate commit messages with a brief justification for the top recommendation.
Use when deciding which Copilot customization primitive to use for a new requirement. Compares instructions, prompts, skills, agents, and hooks and recommends the best fit. Invoke with: /comparar-primitivas, choose primitive, which primitive, instruction vs prompt, skill vs prompt, agent vs skill.
Use when starting exploratory data analysis on a CSV file. Detects columns, types, missing values, and distributions. Proposes initial questions and produces a structured EDA notebook. Invoke with: /csv-eda-basica, eda, explore dataset, analyze csv, check data quality.
Use when documenting a dataset for a project. Creates a structured dataset card covering origin, variables, known biases, licence, and limitations. Invoke with: /dataset-card, document dataset, create dataset card, dataset documentation, data card.
Use when a Python script raises an exception or produces unexpected output. Follows a structured protocol: read the traceback, locate the line, isolate a hypothesis, propose the minimal fix. Invoke with: /debug-python-basico, debug, fix error, traceback, exception, python error.
Use when preparing a complex task for an AI agent. Evaluates what files, examples, or templates should be attached as context to get the best result. Invoke with: /evaluar-contexto-necesario, what context do I need, prepare context, context for agent, attach files.
Use when creating a new Copilot customization file (skill, prompt, agent, or instruction). Proposes the correct frontmatter fields and values. Invoke with: /frontmatter-designer, design frontmatter, write frontmatter, new skill frontmatter, new agent frontmatter.