A Model Context Protocol (MCP) server for GitLab and Jira integration. This server allows AI agents like gemini-cli to interact with your GitLab and Jira instances. - Projects: List all accessible projects or filter them by name. - Merge Requests: List merge requests for a project, get detailed information (including diffs), add comments, and assign reviewers. - Pipeline/CI/CD: Get pipeline status
Add this skill
npx mdskills install HainanZhao/mcp-gitlab-jiraComprehensive GitLab/Jira integration with extensive tooling, clear setup docs, and multi-platform support
A Model Context Protocol (MCP) server for GitLab and Jira integration. This server allows AI agents like gemini-cli to interact with your GitLab and Jira instances.
Install the package globally:
npm i -g mcp-gitlab-jira
Set up environment variables:
# GitLab
export GITLAB_URL="https://your-gitlab-instance.com"
export GITLAB_ACCESS_TOKEN="your-personal-access-token"
# Jira
export ATLASSIAN_SITE_NAME="your-atlassian-site-name"
export ATLASSIAN_USER_EMAIL="your-email@example.com"
export ATLASSIAN_API_TOKEN="your-jira-api-token"
Test the server manually:
# Test that the server starts without errors
echo '{"jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {}}' | mcp-gitlab-jira
The server should start and log "GitLab/Jira MCP server started" to stderr.
Create or update your MCP configuration file (usually ~/.mcp/config.json or similar):
{
"mcpServers": {
"gitlab-jira-mcp": {
"command": "mcp-gitlab-jira",
"env": {
"GITLAB_URL": "https://your-gitlab-instance.com",
"GITLAB_ACCESS_TOKEN": "your-personal-access-token",
"ATLASSIAN_SITE_NAME": "your-atlassian-site-name",
"ATLASSIAN_USER_EMAIL": "your-email@example.com",
"ATLASSIAN_API_TOKEN": "your-jira-api-token"
}
}
}
}
You can also run this MCP server in a Docker container using the pre-built image from Docker Hub.
The Docker images are automatically built and published to Docker Hub for each release:
hainanzhao/mcp-gitlab-jira:latesthainanzhao/mcp-gitlab-jira:v0.1.2, hainanzhao/mcp-gitlab-jira:v0.1.1, etc.The images are built for multiple architectures: linux/amd64 and linux/arm64 (Apple Silicon compatible).
Pull and run the Docker container:
docker run -d --name mcp-gitlab-jira-container \
-e GITLAB_URL="https://your-gitlab-instance.com" \
-e GITLAB_ACCESS_TOKEN="your-personal-access-token" \
-e ATLASSIAN_SITE_NAME="your-atlassian-site-name" \
-e ATLASSIAN_USER_EMAIL="your-email@example.com" \
-e ATLASSIAN_API_TOKEN="your-jira-api-token" \
hainanzhao/mcp-gitlab-jira:latest
Alternative: Run without persistent container (one-time execution):
docker run --rm -i \
-e GITLAB_URL="https://your-gitlab-instance.com" \
-e GITLAB_ACCESS_TOKEN="your-personal-access-token" \
-e ATLASSIAN_SITE_NAME="your-atlassian-site-name" \
-e ATLASSIAN_USER_EMAIL="your-email@example.com" \
-e ATLASSIAN_API_TOKEN="your-jira-api-token" \
hainanzhao/mcp-gitlab-jira:latest
You have two options for using the Docker container with MCP clients:
First, start the container as shown above, then update your MCP configuration file. The env block is empty because the necessary environment variables are passed directly to the container using the -e flag in the docker run command.
{
"mcpServers": {
"gitlab-jira-mcp": {
"command": "docker",
"args": ["exec", "-i", "mcp-gitlab-jira-container", "npm", "start"],
"env": {}
}
}
}
This runs a new container for each MCP session:
{
"mcpServers": {
"gitlab-jira-mcp": {
"command": "docker",
"args": [
"run", "--rm", "-i",
"-e", "GITLAB_URL=https://your-gitlab-instance.com",
"-e", "GITLAB_ACCESS_TOKEN=your-personal-access-token",
"-e", "ATLASSIAN_SITE_NAME=your-atlassian-site-name",
"-e", "ATLASSIAN_USER_EMAIL=your-email@example.com",
"-e", "ATLASSIAN_API_TOKEN=your-jira-api-token",
"hainanzhao/mcp-gitlab-jira:latest"
],
"env": {}
}
}
}
gitlab_get_merge_request_details: Fetches detailed information about a GitLab Merge Request, including file diffs.gitlab_get_file_content: Fetches the content of a specific file at a given SHA in a GitLab project.gitlab_add_comment_to_merge_request: Adds a comment to a GitLab Merge Request. Can be a general comment, a reply to an existing discussion, or an inline comment on a specific line.gitlab_list_merge_requests: Lists merge requests for a given GitLab project.gitlab_assign_reviewers_to_merge_request: Assigns reviewers to a GitLab Merge Request.gitlab_list_project_members: Lists all members (contributors) of a given GitLab project.gitlab_list_project_members_by_project_name: Lists all members (contributors) of a given GitLab project by project name.gitlab_list_projects_by_name: Filters GitLab projects by name using a fuzzy, case-insensitive match.gitlab_list_all_projects: Lists all accessible GitLab projects.gitlab_list_all_releases: Fetches releases for a given GitLab project.gitlab_list_releases_since_version: Filters releases for a given GitLab project since a specific version.gitlab_get_user_id_by_username: Retrieves the GitLab user ID for a given username.gitlab_get_user_activities: Fetches activities for a given GitLab user by their username, optionally filtered by date.gitlab_get_project_pipelines: Gets pipelines for a GitLab project, optionally filtered by branch/ref.gitlab_get_merge_request_pipelines: Gets pipelines for a specific GitLab Merge Request.gitlab_get_pipeline_details: Gets detailed information about a specific pipeline.gitlab_get_pipeline_jobs: Gets jobs for a specific pipeline.gitlab_get_job_logs: Gets logs for a specific job.gitlab_trigger_pipeline: Triggers a new pipeline for a specific branch/ref.gitlab_retry_pipeline: Retries a failed pipeline.gitlab_cancel_pipeline: Cancels a running pipeline.gitlab_list_branches: Lists all branches in a GitLab project.gitlab_create_branch: Creates a new branch in a GitLab project.gitlab_delete_branch: Deletes a branch from a GitLab project.gitlab_get_branch_details: Gets detailed information about a specific branch.gitlab_list_project_issues: Lists issues in a GitLab project.gitlab_get_issue_details: Gets detailed information about a specific GitLab issue.gitlab_create_issue: Creates a new issue in a GitLab project.gitlab_update_issue: Updates an existing GitLab issue.gitlab_close_issue: Closes a GitLab issue.gitlab_add_comment_to_issue: Adds a comment to a GitLab issue.gitlab_get_issue_comments: Gets comments for a GitLab issue.jira_get_ticket_details: Fetches comprehensive information about a Jira ticket with flattened fields, including all custom fields with user-friendly names. Automatically filters out empty values and less useful fields (attachments, avatars). Returns both system and custom fields in a clean, flat structure.jira_get_ticket_comments: Fetches comments for a Jira ticket.jira_add_comment_to_ticket: Adds a comment to a Jira ticket.jira_search_tickets_by_jql: Searches for Jira tickets using a JQL (Jira Query Language) string.jira_create_ticket: Creates a new Jira ticket with given fields.jira_get_available_transitions: Fetches available transitions for a Jira ticket.jira_update_ticket: Updates a Jira ticket summary, description, labels.jira_update_custom_fields: Updates custom fields on a Jira ticket.jira_update_ticket_priority: Updates the priority value for a Jira ticket. Automatically finds the priority custom field (case-insensitive search for "Priority" or "priority"), fetches the allowed values for that field, and matches the provided priority name using fuzzy string matching. Accepts values like "Critical", "High", "Medium", "Low", etc., and will find the best match from the predefined options.jira_update_ticket_sprint: Updates the sprint value for a Jira ticket. Automatically finds the sprint custom field (case-insensitive search for "Sprint" or "sprint"), fetches available sprints from the project's boards, and matches the provided sprint name using fuzzy string matching. Accepts values like "Sprint 1", "Bug Fix Sprint", etc., and will find the best match from active and future sprints in the project.jira_transition_ticket: Transitions a Jira ticket to a new status.jira_get_all_fields: Fetches the list of all fields from Jira.jira_get_all_projects: Gets all accessible Jira projects.jira_get_project_details: Gets detailed information about a specific Jira project.jira_get_project_components: Gets components for a Jira project.jira_get_project_versions: Gets versions for a Jira project.npm install and npm run build.GITLAB_ACCESS_TOKEN, ATLASSIAN_USER_EMAIL, and ATLASSIAN_API_TOKEN have the necessary permissions.GITLAB_URL and ATLASSIAN_SITE_NAME are correct and accessible.To see detailed logs, you can run the server directly:
export GITLAB_URL="your-url"
export GITLAB_ACCESS_TOKEN="your-token"
export ATLASSIAN_SITE_NAME="your-atlassian-site-name"
export ATLASSIAN_USER_EMAIL="your-email@example.com"
export ATLASSIAN_API_TOKEN="your-jira-api-token"
mcp-gitlab-jira
For development, clone the repository and install the dependencies.
npm install
npm run build
To test the Docker build locally before pushing:
# Build and test the Docker image locally
./scripts/build-docker-local.sh
This script will build the Docker image and run basic tests to ensure it works correctly.
For maintainers: See Docker Setup Guide for information about setting up automated Docker Hub publishing.
src/index.ts: Main MCP server implementationsrc/gitlab.service.ts: GitLab API clientsrc/gitlab.ts: GitLab type definitionssrc/jira.service.ts: Jira API clientsrc/jira.ts: Jira type definitionsdist/: Compiled JavaScript outputGitLabService or JiraService class.allTools array in index.ts.index.ts.npm run build.ISC
Install via CLI
npx mdskills install HainanZhao/mcp-gitlab-jiraMCP GitLab Jira Server is a free, open-source AI agent skill. A Model Context Protocol (MCP) server for GitLab and Jira integration. This server allows AI agents like gemini-cli to interact with your GitLab and Jira instances. - Projects: List all accessible projects or filter them by name. - Merge Requests: List merge requests for a project, get detailed information (including diffs), add comments, and assign reviewers. - Pipeline/CI/CD: Get pipeline status
Install MCP GitLab Jira Server with a single command:
npx mdskills install HainanZhao/mcp-gitlab-jiraThis downloads the skill files into your project and your AI agent picks them up automatically.
MCP GitLab Jira Server works with Claude Code, Claude Desktop, Cursor, Vscode Copilot, Windsurf, Continue Dev, Gemini Cli, Amp, Roo Code, Goose. Skills use the open SKILL.md format which is compatible with any AI coding agent that reads markdown instructions.