Token导航 LogoToken导航TokenDH.com
待分类external-servicegithub未标认证来源可访问许可证需确认审计通过

skill-improvement技能提升

Agent Skill

skill-improvement 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

306

周安装

13

GitHub Stars

公开资料未说明

下载量

107
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:skill-improvement(技能提升)
来源仓库:https://github.com/diegocanepa/agent-skills
仓库路径:skills/skill-improvement
安装命令:
npx skills add https://github.com/diegocanepa/agent-skills --skill skill-improvement
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/diegocanepa/agent-skills --skill skill-improvement

简介

skill-improvement 用于识别现有技能的不足并提出优化方案。

  • 适用于发现缺失步骤、边缘情况或指令模糊的技能改进。
  • 可建议新增引用其他技能或 MCP 服务以增强能力。
  • 需明确目标技能和具体改进点后方可提交更新提案。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

🔧 Skill Improvement

Identify gaps or enhancement opportunities in an existing skill from the current conversation and contribute an improved version to the shared repository.

When to Use

Use this skill when:

  • A skill didn't behave as expected and you've identified the root cause.
  • The conversation reveals a missing step, edge case, or unclear instruction in a skill.
  • A skill should reference another skill or MCP that wasn't mentioned.
  • The user says "improve this skill", "update this skill", or "this skill needs X".

Workflow

Step 1: Identify the Target Skill

Clarify with the user (if not already clear):

  • Which skill needs improvement? (e.g. swe-orchestrator, vcs-commit)
  • What specifically should be improved? (missing step, wrong reference, unclear instruction, new edge case)

Step 2: Read the Current Skill

Read the current SKILL.md of the target skill from skills/<skill-name>/SKILL.md.

Step 3: Extract Improvements from the Conversation

Review the conversation and identify:

  • What failed or was unclear in the skill's current instructions.
  • What new patterns, steps, or guardrails should be added.
  • Whether any references to other skills, MCPs, or tools are missing.
  • Whether the description frontmatter still accurately describes when to use the skill.

Present a diff summary to the user before making changes:

Proposed Improvements:
- [ ] [Section/Line]: [What changes and why]
- [ ] [Section/Line]: [What changes and why]

Step 4: Apply the Improvements

Edit the skills/<skill-name>/SKILL.md file with the identified improvements.

Quality check after editing:

Improvement Review:
- [ ] The `description` frontmatter still accurately triggers the skill
- [ ] New instructions are self-contained and clear
- [ ] No existing correct behavior was accidentally removed
- [ ] References to other skills use relative paths (e.g. [vcs-commit](../vcs-commit/SKILL.md))
- [ ] Change is focused — does not introduce unrelated refactoring

If any check fails → revise → re-check.

Step 5: Create the Branch

Use vcs-branch to create a dedicated branch:

fix/improve-skill-<skill-name>

or for larger additions:

feat/enhance-skill-<skill-name>

Step 6: Commit the Changes

Use vcs-commit to commit:

git add skills/<skill-name>/SKILL.md
git commit -m "fix(skills): improve <skill-name> - <short description of what changed>"
git push -u origin fix/improve-skill-<skill-name>

Step 7: Create the Pull Request

Use vcs-change-request and the GitHub MCP to open a PR:

  • Owner: diegocanepa
  • Repo: agent-skills
  • URL: https://github.com/diegocanepa/agent-skills
  • Head branch: fix/improve-skill-<skill-name>
  • Base branch: main

PR description template:

## Summary

Improves the **<skill-name>** skill.

## Problem

<What was unclear, missing, or broken in the current version.>

## Changes

- [Section changed]: <What was added/modified and why>
- [Section changed]: <What was added/modified and why>

## Conversation Context

This improvement was identified from a conversation about: <topic>.

## Checklist
- [ ] Existing correct behavior preserved
- [ ] Instructions are clearer and more actionable
- [ ] No unrelated changes introduced

Step 8: MITM — Confirm with User

"I've applied the improvements to <skill-name> and the PR is ready at https://github.com/diegocanepa/agent-skills. Does the diff look correct?"

Feedback Loop: If the USER wants to refine further → adjust → amend/push → re-present.


Notes

  • Keep improvements focused. One PR per skill, one concern per PR.
  • Never rewrite a skill entirely unless the whole structure is broken. Prefer targeted edits.
  • If the improvement is substantial and changes the skill's purpose, consider using add-skill to create a new variant instead.
  • The description frontmatter is critical — if the improvement changes *when* the skill applies, update it.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.93%
按下载量换算36

Claude

28.49%
按下载量换算30

Cursor

18.61%
按下载量换算20

Gemini CLI

9.21%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills