Token导航 LogoToken导航TokenDH.com
研究检索操作浏览器github未标认证来源可访问clear审计提醒

gitlabGitLab 代码协作

Agent Skill

用于围绕 GitLab 项目、Merge Request、Issue、分支、流水线和代码审查流程提供辅助能力。它适合让 Agent 查询项目状态、整理提交差异、辅助检查合并请求或汇总 CI 结果。使用时需要确认项目权限、访问 token 和目标分支范围;涉及合并、推送、改工单或触发流水线时,应先预览影响并核对团队流程。

总安装

412

周安装

17

GitHub Stars

175

下载量

135
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

3

许可证

MIT

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:gitlab(GitLab 代码协作)
来源仓库:https://github.com/elithrar/dotfiles
仓库路径:skills/gitlab
安装命令:
npx skills add https://github.com/elithrar/dotfiles --skill gitlab
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/elithrar/dotfiles --skill gitlab

简介

gitlab 用于围绕 GitLab 项目、Merge Request 和 Issue 提供辅助能力。

  • 适合让 Agent 查询项目状态、整理提交差异或汇总 CI 结果。
  • 使用时需要确认项目权限和访问 token;涉及合并、推送或改工单时应先预览影响并核对团队流程。
  • 可通过 npx skills add 命令从指定 GitHub 仓库安装,建议确认权限和维护状态后再使用。
  • gitlab 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

GitLab Workflow

Operate on GitLab repositories using the glab CLI — merge requests, CI/CD pipelines, issues, and releases.

FIRST: Verify Environment

Run both checks before any glab command. If either fails, STOP.

git remote -v | grep -i gitlab
glab auth status

If glab is not installed, ask the user to install it (brew install glab on macOS, sudo apt install glab on Debian/Ubuntu). If not authenticated, run glab auth login. For self-managed instances, use glab auth login --hostname gitlab.example.com.

Critical Rules

  • Avoid interactive commands. glab ci view launches a TUI the agent cannot operate. Use glab ci status for pipeline state and glab ci view --web to open in browser. Always pass --fill --yes to glab mr create to skip interactive prompts.
  • Check CI before requesting review. Run glab ci status after pushing. A broken pipeline wastes reviewer time and signals the MR isn't ready.
  • Pull logs first on CI failure. Run glab ci trace <job-name> before anything else — 90% of failures are self-explanatory from the logs.

Behavioral Guidelines

  • Use draft MRs for work in progress: create with --draft, mark ready with glab mr update <id> --ready
  • Reference issues explicitly: Closes #123 for auto-close, Relates to #123 for tracking only
  • Retry flaky failures, fix real ones: glab ci retry <job-name> once. If it fails again, it's a real issue — read the logs and fix the code.

Quick Reference: MR Commands

TaskCommand
Create MR (non-interactive)glab mr create --fill --yes -b main
Create draft MRglab mr create --fill --yes --draft -b main
List open MRsglab mr list
View MR detailsglab mr view 123
View MR commentsglab mr view 123 --comments
View MR diffglab mr diff 123
Check out MR locallyglab mr checkout 123
Mark draft as readyglab mr update 123 --ready
Add reviewersglab mr update 123 --reviewer user1,user2
Approve MRglab mr approve 123
Squash and mergeglab mr merge 123 --squash
Merge without waiting for CIglab mr merge 123 --auto-merge=false
Add MR commentglab mr note 123 -m "message"

Quick Reference: CI Commands

TaskCommand
Check pipeline statusglab ci status
Watch pipeline liveglab ci status --live
Open pipeline in browserglab ci view --web
List recent pipelinesglab ci list
Pull job logsglab ci trace <job-name>
Retry failed jobglab ci retry <job-name>
Run new pipelineglab ci run
Run with variablesglab ci run --variables "KEY:value"
Download artifactsglab job artifact <ref> <job-name>

Workflow: Branch to Merged MR

After branching, committing, and pushing:

glab mr create --fill --yes -b main    # create MR
glab ci status --live                   # watch pipeline
# if CI fails:
glab ci trace <failed-job-name>         # pull logs, fix, push
glab ci status --live                   # verify fix
# when CI passes and MR is approved:
glab mr merge --squash --remove-source-branch

CI Failure Remediation

  1. glab ci status — identify which pipeline failed
  2. glab ci trace <job-name> — pull the failed job's logs
  3. Read the logs from the bottom up — the final error is usually the root cause
  4. Reproduce locally if possible (run the same commands from .gitlab-ci.yml)
  5. Fix, commit, push
  6. glab ci status --live — watch the new pipeline
  7. If still failing, repeat from step 2

Common CI Log Signatures

PatternLog Signature
Test failureFAIL, AssertionError, expected... got
Build failurecompilation failed, cannot find module
TimeoutJob timed out, exceeded limit
Missing env/secretundefined, authentication failed
Dependency issue404 Not Found, version not found

For operations not covered by glab subcommands, use glab api directly. For detailed command options, issue management, releases, search, and advanced CI debugging, see references/commands.md.

Use --fill --yes for MR creation and glab ci status instead of glab ci view — interactive TUI commands will hang.

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

OpenCode

30.21%
按下载量换算41

Claude Code

21.93%
按下载量换算30

windsurf

14.8%
按下载量换算20

trae

12.39%
按下载量换算17

Codex

7.4%
按下载量换算10

Antigravity

3.41%
按下载量换算5

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。

来源信息

继续浏览同类 Skills