GitLab CI Pipeline Architect
Design, debug, and optimize GitLab CI/CD pipelines with smart decision trees
What You Can Do
This skill helps you architect robust GitLab CI/CD pipelines by analyzing your pipeline structure, identifying bottlenecks, and recommending optimizations. You get actionable debugging strategies for failing jobs, performance analysis with concrete improvement suggestions, and decision trees for complex pipeline design scenarios. Whether you're scaling pipelines for enterprise workloads or fixing flaky deployments, this skill delivers pipeline expertise without needing a DevOps consultant.
Features
Evaluate your .gitlab-ci.yml structure and recommend improvements for clarity, maintainability, and performance
Systematically diagnose failing jobs with root cause analysis, log patterns, and targeted fixes
Identify slow stages, inefficient caching, and parallel execution opportunities
Refactor CI scripts for readability, DRY principles, and GitLab best practices
Navigate complex scenarios like multi-environment deployments, artifact management, and branch strategies
Learn from pipeline failures to prevent regressions and improve reliability
Reduce runner minutes and infrastructure costs through smarter job scheduling and resource allocation
Example Output
Job Failure Analysis:
-
❌ Problem:
build-servicejob failing with Docker auth errors -
✅ Root Cause:
DOCKER_AUTH_CONFIGvariable not passed to job -
✅ Solution:
build-service:
script:
- docker build -t myapp:latest .
variables:
DOCKER_AUTH_CONFIG: $DOCKER_AUTH_CONFIG
Performance Optimization Report:
- 📊 Current pipeline: 45 minutes
- 🎯 Optimized pipeline: 18 minutes (60% faster)
Changes:
- Parallelize test jobs (4 → 12 concurrent jobs)
- Add Docker layer caching with
DOCKER_TLS_CERTDIR - Split artifact upload to background job
Expected savings: 27 minutes per pipeline × 50 runs/week = 22.5 hours/week
What's Included
- SKILL.md: Complete pipeline debugging and optimization workflows
- GitLab CI troubleshooting decision tree: Step-by-step flowchart for common failures
- Pipeline performance checklist: Audit your setup against 25+ optimization criteria
- YAML configuration templates: Docker builds, multi-stage deployments, artifact management
- Job failure pattern guide: Common errors and proven fixes
- Multi-environment workflow: Best practices for dev/staging/prod promotion pipelines
Who It's For
- DevOps engineers building and scaling CI/CD infrastructure
- Platform engineers managing multi-team GitLab deployments
- Software engineers debugging flaky or failing pipeline tests
- Infrastructure teams optimizing runner resource utilization and costs
- Engineering leads modernizing legacy CI/CD setups
Best For
- Diagnosing and fixing failing GitLab CI jobs and stages
- Redesigning pipelines for better performance and maintainability
- Optimizing runner costs and total execution time
- Implementing best practices for complex multi-environment deployments
- Creating reusable pipeline templates and patterns for teams
You might also like
Transform breaking news briefs into compelling visual story frameworks that guide photographers toward impactful coverage. You'll receive multiple narrative angles, detailed shot lists organized by scene and purpose, and complete assignment briefs with sourcing guidance. Each output ensures comprehensive emotional and contextual storytelling from first frame to final edit.
by manwithstyle
You'll receive data-driven recommendations for chart types, color schemes, visual hierarchies, and narrative structures tailored to your audience and message. This skill guides you through selecting encodings that accurately represent data while maximizing audience comprehension and engagement—whether you're designing publication-ready graphics or interactive dashboards.
by manwithstyle
You can rapidly evaluate public datasets to uncover newsworthy patterns and develop story angles backed by reproducible analysis. Claude helps you assess data quality, identify anomalies, generate multiple narrative angles, and document your methodology so editors and fact-checkers can verify your findings. Turn raw data into compelling stories faster than traditional research.
by manwithstyle
Quickly architect scalable, reusable Terraform modules that follow HashiCorp best practices and organizational standards. Get automated reviews that catch common pitfalls—variable naming, provider configuration, resource dependencies—before they reach production. Standardize your infrastructure-as-code across teams with instant feedback on module quality, security posture, and cost optimization opportunities.
by manwithstyle
Analyze Docker container logs instantly to identify root causes of failures, crashes, and resource bottlenecks. Get actionable recommendations for resource allocation, network configuration, and architectural improvements that eliminate production issues before they cascade.
by manwithstyle
Manage your entire story development pipeline—from assignment briefs and source research guidance to fact-checking verification and copy editing—all within Claude's context. You'll generate assignment templates, receive real-time editorial feedback, identify verification gaps, and receive copyediting suggestions with tracked changes. The skill handles complex multi-source stories, deadline pressure, and maintains editorial standards across your publication.
by manwithstyle
You can transform complex datasets into compelling interactive narratives that maintain coherence while inviting exploration. This skill helps you architect data-driven stories where insights unfold naturally through guided discovery points, keeping readers engaged without limiting their agency. Readers navigate your narrative with purpose, uncovering patterns and relationships that would remain hidden in static presentations.
by manwithstyle
Rapidly triage infrastructure incidents across all GCP services using guided diagnostic workflows, root cause decision trees, and remediation procedures. You get structured steps to isolate failures in Compute Engine, Cloud Run, Cloud SQL, networking, and storage, complete with gcloud commands, health checks, and rollback procedures.
by manwithstyle