Skip to content

POWER.md

Latest commit

 

History

History
123 lines (94 loc) · 4.79 KB

POWER.md

File metadata and controls

123 lines (94 loc) · 4.79 KB
Raw

CircleCI CI/CD Pipeline Management Power

A Kiro power that integrates the official CircleCI MCP Server for managing CI/CD pipelines directly from your IDE. Includes best-practice steering files for Python data engineering projects.

What This Power Does

  • Validate & Fix Configs — Validate .circleci/config.yml against CircleCI schema and get fix suggestions
  • Debug Build Failures — Retrieve failure logs and diagnose issues without leaving the IDE
  • Run Pipelines — Trigger pipelines, rerun workflows, or rollback deployments
  • Monitor Status — Check latest pipeline status for any branch
  • Find Flaky Tests — Identify unreliable tests from execution history
  • Optimize Resources — Find jobs with underused compute resource classes
  • Best Practices — Steering files with CircleCI patterns for Python, OIDC, CDK, and Glue

Prerequisites

  1. CircleCI API Token — Generate a personal API token at https://app.circleci.com/settings/user/tokens
  2. Node.js — Required to run the MCP server via npx
  3. Environment Variable — Set CIRCLECI_TOKEN in your environment or .env file

Setup

1. Set your CircleCI token

Add to your shell profile or .env:

export CIRCLECI_TOKEN="your-circleci-personal-api-token"

2. Install the power

Copy this directory to your Kiro powers location, or reference it in your workspace MCP config.

3. MCP Configuration

The power registers the official CircleCI MCP server. If you need to configure it manually, add to .kiro/settings/mcp.json:

{
  "mcpServers": {
    "circleci": {
      "command": "npx",
      "args": ["-y", "@circleci/mcp-server-circleci"],
      "env": {
        "CIRCLECI_TOKEN": "${CIRCLECI_TOKEN}",
        "CIRCLECI_BASE_URL": "https://circleci.com"
      },
      "disabled": false,
      "autoApprove": ["list_followed_projects", "get_latest_pipeline_status", "config_helper"]
    }
  }
}

Available Tools (from Official MCP Server)

Tool Description
config_helper Validate and fix CircleCI configuration YAML
get_build_failure_logs Retrieve detailed failure logs from builds
get_job_test_results Get test metadata and results for jobs
get_latest_pipeline_status Check pipeline status for a branch
find_flaky_tests Identify flaky tests from execution history
find_underused_resource_classes Find jobs with underused compute
list_followed_projects List all projects you follow
run_pipeline Trigger a pipeline to run
rerun_workflow Rerun a workflow from start or from failed job
run_rollback_pipeline Trigger a rollback for a project
list_artifacts List artifacts produced by a job
analyze_diff Analyze git diffs against rules for violations

What's Included

kiro-powers-circleci/
├── POWER.md                          # This file
├── package.json                      # Power metadata + MCP server config
├── steering/
│   ├── circleci-best-practices.md    # General CircleCI best practices
│   ├── python-pipeline-patterns.md   # Python/PySpark CI patterns
│   └── oidc-aws-patterns.md          # OIDC + AWS deployment patterns
├── hooks/
│   ├── validate-config-on-edit.json  # Auto-validate config on save
│   └── check-pipeline-on-push.json  # Check pipeline status after push
└── templates/
    ├── python-data-eng.yml           # Template: Python data engineering pipeline
    ├── cdk-deploy.yml                # Template: CDK synth + deploy pipeline
    └── glue-workflow.yml             # Template: Glue job upload workflow

Usage Examples

Validate your config

"Validate my CircleCI config and fix any issues"

Debug a failure

"Get the build failure logs for my latest pipeline on the main branch"

Check status

"What's the status of my latest pipeline?"

Run a pipeline

"Trigger a pipeline run on my feature branch"

Find flaky tests

"Find flaky tests in my project"

Integration with AIDLC

When used alongside kiro-powers-aidlc, this power is activated during:

  • Build and Test stage — Validates generated CircleCI configs
  • Code Generation — Provides pipeline templates for new services
  • Post-deployment — Monitors pipeline status after code changes

Steering Files

The steering files provide context-aware guidance:

  • circleci-best-practices.md — Caching, parallelism, orbs, resource classes, security
  • python-pipeline-patterns.md — pytest, coverage, linting, PySpark testing patterns
  • oidc-aws-patterns.md — OIDC federation, CDK deploy, multi-environment promotion

These are automatically included when working on .circleci/ files or CI/CD tasks.