CircleCI

CircleCI

Enable AI Agents to fix build failures from CircleCI.

69
Stars
39
Forks
0
Releases

Overview

CircleCI MCP Server provides an MCP interface for CircleCI, enabling natural-language interactions with CircleCI through MCP clients such as Cursor IDE, Windsurf, Copilot, and other MCP-enabled tools. It translates user requests into CircleCI API calls, enabling tasks like finding the latest failed pipeline and retrieving logs, or triggering pipelines, checking statuses, retrieving test results, and more—without leaving your editor. The server supports three input modalities: by project slug and branch, by direct CircleCI URLs (pipeline or job), or by local project context (workspace path, git remote, branch). Features are exposed as tools (e.g., get_build_failure_logs, find_flaky_tests, get_latest_pipeline_status, get_job_test_results, config_helper, create_prompt_template, list_followed_projects, run_pipeline). It requires a CircleCI Personal API Token and Node.js ≥18. Installation options include NPX, Docker, or a self-managed remote MCP server, and it includes guides for integration with Cursor, VS Code, Claude Desktop/Code, and Windsurf. Development and troubleshooting guidance are provided in the repo.

Details

Owner
CircleCI-Public
Language
TypeScript
License
Other
Updated
2025-12-07

Features

get_build_failure_logs

Retrieves detailed failure logs from CircleCI builds; supports input via project slug/branch, CircleCI URLs, or local project context; outputs job names, step details, and failure context.

find_flaky_tests

Identifies flaky tests in a CircleCI project by analyzing test history; supports input via project slug, CircleCI URL, or local project context; returns flaky tests with details.

get_latest_pipeline_status

Retrieves the status of the latest pipeline for a given branch; supports input via project slug/branch, CircleCI URL, or local project context; returns workflow statuses, durations, timestamps, and overall health.

get_job_test_results

Retrieves test metadata for CircleCI jobs; provides a summary of tests, details of failed tests with location and messages, and duration information.

config_helper

Assists with CircleCI configuration tasks by validating config.yml and providing validation results and recommendations.

create_prompt_template

Generates structured prompt templates for AI-enabled applications from feature requirements; returns a prompt template and a context schema.

list_followed_projects

Lists all CircleCI projects the user is following; provides project names and projectSlug for use in subsequent tool calls.

run_pipeline

Triggers a CircleCI pipeline; supports input via project slug/branch, CircleCI URLs, or local project context; returns a monitor link.

Tags

CircleCIMCPMCP ServerAI AgentsCursor IDEWindsurfCopilotVS CodeClaude DesktopDockerNPXSelf-Managed Remote