Agent skill

glab

Expert guidance for using the GitLab CLI (glab) to manage GitLab issues, merge requests, CI/CD pipelines, repositories, and other GitLab operations from the command line. Use this skill when the user needs to interact with GitLab resources or perform GitLab workflows.

Stars 114
Forks 19

Install this agent skill to your Project

npx add-skill https://github.com/NikiforovAll/claude-code-rules/tree/main/plugins/handbook-glab/skills/glab-skill

SKILL.md

GitLab CLI (glab) Skill

Provides guidance for using glab, the official GitLab CLI, to perform GitLab operations from the terminal.

When to Use This Skill

Invoke when the user needs to:

  • Create, review, or manage merge requests
  • Work with GitLab issues
  • Monitor or trigger CI/CD pipelines
  • Clone or manage repositories
  • Perform any GitLab operation from the command line

Prerequisites

Verify glab installation before executing commands:

bash
glab --version

If not installed, inform the user and provide platform-specific installation guidance.

Authentication Quick Start

Most glab operations require authentication:

bash
# Interactive authentication
glab auth login

# Check authentication status
glab auth status

# For self-hosted GitLab
glab auth login --hostname gitlab.example.org

# Using environment variables
export GITLAB_TOKEN=your-token
export GITLAB_HOST=gitlab.example.org  # for self-hosted

Core Workflows

Creating a Merge Request

bash
# 1. Ensure branch is pushed
git push -u origin feature-branch

# 2. Create MR
glab mr create --title "Add feature" --description "Implements X"

# With reviewers and labels
glab mr create --title "Fix bug" --reviewer=alice,bob --label="bug,urgent"

Reviewing Merge Requests

bash
# 1. List MRs awaiting your review
glab mr list --reviewer=@me

# 2. Checkout MR locally to test
glab mr checkout <mr-number>

# 3. After testing, approve
glab mr approve <mr-number>

# 4. Add a general comment
glab mr note <mr-number> -m "Please update tests"

Posting Inline DiffNote Comments

To anchor a comment to a specific line in the diff (a DiffNote), you must use a JSON body with the Content-Type: application/json header. Using -F form fields will create an unanchored DiscussionNote.

bash
# Get SHA values from the MR
glab mr view <id> --output=json | grep -oP '"(base_sha|head_sha|start_sha)":\s*"\K[0-9a-f]{40}'

# Post a DiffNote on a new line
glab api --method POST "projects/:id/merge_requests/:iid/discussions" \
  -H "Content-Type: application/json" \
  --input - <<'JSONEOF'
{
  "body": "Your review comment here",
  "position": {
    "base_sha": "<base_sha>",
    "start_sha": "<start_sha>",
    "head_sha": "<head_sha>",
    "position_type": "text",
    "new_path": "path/to/file.js",
    "new_line": 42
  }
}
JSONEOF

Key rules for DiffNotes:

  • Always use --input - with a heredoc and -H "Content-Type: application/json". Never use -F for the position object — it will silently create a DiscussionNote instead.
  • Line numbers are new file line numbers calculated from the diff hunk headers (@@ -old,count +new,count @@), not the diff output line numbers. Count only context lines and + lines from the hunk start.
  • Verify findings on the MR source branch (git fetch origin <branch> then git grep), not the current working branch.
  • To update an existing note: glab api --method PUT "projects/:id/merge_requests/:iid/notes/:note_id" -H "Content-Type: application/json" --input - <<< '{"body":"updated text"}'
  • GitLab has no batch "Submit Review" API like GitHub. Post DiffNotes individually.
  • To find a discussion or note ID: glab api "projects/:id/merge_requests/:iid/discussions" --paginate

Managing Issues

bash
# Create issue with labels
glab issue create --title "Bug in login" --label=bug

# Link MR to issue
glab mr create --title "Fix login" --description "Closes #<issue-number>"

# List your assigned issues
glab issue list --assignee=@me

Monitoring CI/CD

bash
# Watch pipeline in progress
glab pipeline ci view

# Check pipeline status
glab ci status

# View logs if failed
glab ci trace

# Retry failed pipeline
glab ci retry

# Lint CI config before pushing
glab ci lint

Common Patterns

Working Outside Repository Context

When not in a Git repository, specify the repository:

bash
glab mr list -R owner/repo
glab issue list -R owner/repo

Self-Hosted GitLab

Set hostname for all commands:

bash
export GITLAB_HOST=gitlab.example.org
# or per-command
glab repo clone gitlab.example.org/owner/repo

Listing Unresolved MR Comments

bash
glab api "projects/:id/merge_requests/{mr}/discussions?per_page=100" | jq '[.[] | select(.notes[0].resolvable == true and .notes[0].resolved == false) | {id: .notes[0].id, body: .notes[0].body[0:100], path: .notes[0].position.new_path, line: .notes[0].position.new_line}]'

Automation and Scripting

Use JSON output for parsing:

bash
glab mr list --output=json | jq '.[] | .title'

Replying to MR Notes/Threads

glab mr note creates standalone comments. To reply within a discussion thread, use the API:

bash
# 1. Find the discussion_id containing the note
glab api "projects/:id/merge_requests/{mr}/discussions" | jq '.[] | select(.notes[].id == {note_id}) | .id'

# 2. Post reply to the discussion thread
glab api --method POST "projects/:id/merge_requests/{mr}/discussions/{discussion_id}/notes" --field body="Your reply"

Example:

bash
# Get discussion_id for note 13698970
glab api "projects/:id/merge_requests/1013/discussions" | jq '.[] | select(.notes[].id == 13698970) | {id}'
# Returns: {"id": "5356c3552e72e7b4c49276eb4dacfe3efe5c2c5c"}

# Reply to that thread
glab api --method POST "projects/:id/merge_requests/1013/discussions/5356c3552e72e7b4c49276eb4dacfe3efe5c2c5c/notes" --field body="Thanks for the review!"

Using the API Command

The glab api command provides direct GitLab API access:

bash
# Basic API call
glab api projects/:id/merge_requests

# IMPORTANT: Pagination uses query parameters in URL, NOT flags
# ❌ WRONG: glab api --per-page=100 projects/:id/jobs
# ✓ CORRECT: glab api "projects/:id/jobs?per_page=100"

# Auto-fetch all pages
glab api --paginate "projects/:id/pipelines/123/jobs?per_page=100"

# POST with data
glab api --method POST projects/:id/issues --field title="Bug" --field description="Details"

Best Practices

  1. Verify authentication before executing commands: glab auth status
  2. Use --help to explore command options: glab <command> --help
  3. Link MRs to issues using "Closes #123" in MR description
  4. Lint CI config before pushing: glab ci lint
  5. Check repository context when commands fail: git remote -v

Common Commands Quick Reference

Merge Requests:

  • glab mr list --assignee=@me - Your assigned MRs
  • glab mr list --reviewer=@me - MRs for you to review
  • glab mr create - Create new MR
  • glab mr checkout <number> - Test MR locally
  • glab mr approve <number> - Approve MR
  • glab mr merge <number> - Merge approved MR

Issues:

  • glab issue list - List all issues
  • glab issue create - Create new issue
  • glab issue close <number> - Close issue

CI/CD:

  • glab pipeline ci view - Watch pipeline
  • glab ci status - Check status
  • glab ci lint - Validate .gitlab-ci.yml
  • glab ci retry - Retry failed pipeline

Repository:

  • glab repo clone owner/repo - Clone repository
  • glab repo view - View repo details
  • glab repo fork - Fork repository

Progressive Disclosure

For detailed command documentation, refer to:

  • references/commands-detailed.md - Comprehensive command reference with all flags and options
  • references/quick-reference.md - Condensed command cheat sheet
  • references/troubleshooting.md - Detailed error scenarios and solutions

Load these references when:

  • User needs specific flag or option details
  • Troubleshooting authentication or connection issues
  • Working with advanced features (API, schedules, variables, etc.)

Common Issues Quick Fixes

"command not found: glab" - Install glab or verify PATH

"401 Unauthorized" - Run glab auth login

"404 Project Not Found" - Verify repository name and access permissions

"not a git repository" - Navigate to repo or use -R owner/repo flag

"source branch already has a merge request" - Use glab mr list to find existing MR

For detailed troubleshooting, load references/troubleshooting.md.

Notes

  • glab auto-detects repository context from Git remote
  • Most commands have --web flag to open in browser
  • Use --output=json for scripting and automation
  • Multiple GitLab accounts can be authenticated simultaneously
  • Commands respect Git configuration and current repository context

Expand your agent's capabilities with these related and highly-rated skills.

NikiforovAll/claude-code-rules

update-component-reference

This skill should be used when the user wants to add components (commands, agents, skills, hooks, or MCP servers) to the Component Reference section of the website.

114 19
Explore
NikiforovAll/claude-code-rules

version-bump

This skill automates version bumping during the release process for the Claude Code Handbook monorepo. It should be used when the user requests to bump versions, prepare a release, or increment version numbers across the repository.

114 19
Explore
NikiforovAll/claude-code-rules

spec-driven

Guide spec-driven development workflow (Requirements → Design → Tasks → Implementation) with approval gates between phases. Use when user wants structured feature planning or says "use spec-driven" or "follow the spec process".

114 19
Explore
NikiforovAll/claude-code-rules

nano-banana-prompting

This skill should be used when crafting prompts for Nano Banana Pro (Gemini image generation). Use when users want help writing image generation prompts, need guidance on prompt structure, or want to optimize their prompts for better results.

114 19
Explore
NikiforovAll/claude-code-rules

nano-banana

This skill should be used for Python scripting and Gemini image generation. Use when users ask to generate images, create AI art, edit images with AI, or run Python scripts with uv. Trigger phrases include "generate an image", "create a picture", "draw", "make an image of", "nano banana", or any image generation request.

114 19
Explore
NikiforovAll/claude-code-rules

structured-plan-mode

This skill should be used when planning and tracking complex feature implementations that require systematic task decomposition. Use this skill to break down large features into manageable, well-documented tasks with clear dependencies, action items, and success criteria. The skill provides a structured template and methodology for iterative planning and tracking throughout implementation.

114 19
Explore

Didn't find tool you were looking for?

Be as detailed as possible for better results