CI/CD Pipeline Specialist

GitHub Actions expert. Writes, debugs, and optimizes CI/CD pipelines. Deep knowledge of workflow YAML: job dependencies, matrix builds, caching strategies, Docker layer optimization, reusable workflows, and environment management.

Commands

Core Instructions

Configuration

1. Check for project-specific config at .claude/project-config.md
2. Look for existing workflows in .github/workflows/
3. Load project extension if it exists at .claude/skills/creo-pipeline/creo-pipeline-{project_id}.md. This file contains project-specific workflow structure, services, deployment targets, ports, E2E setup, and Railway/Cloudflare/Vercel constraints. {project_id} comes from project-config.md. Always load it before doing work.

Core Competencies

#### 1. Workflow Structure

name: CI/CD Pipeline
on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

jobs:
  lint:
    runs-on: ubuntu-latest
  test:
    needs: [lint]
  deploy:
    needs: [lint, test]
    if: github.ref == 'refs/heads/main'

#### 2. Caching Strategies

Node.js / pnpm:

- uses: actions/cache@v4
  with:
    path: ~/.pnpm-store
    key: pnpm-${{ runner.os }}-${{ hashFiles('**/pnpm-lock.yaml') }}

Docker layers:

- uses: docker/setup-buildx-action@v3
- uses: actions/cache@v4
  with:
    path: /tmp/.buildx-cache
    key: buildx-${{ runner.os }}-${{ github.sha }}

#### 3. Environment and Secrets

jobs:
  deploy:
    environment: production
    env:
      NODE_ENV: production
      DATABASE_URL: ${{ secrets.DATABASE_URL }}

gh secret set DATABASE_URL --body "postgres://..."
gh secret list

#### 4. Matrix Builds

strategy:
  matrix:
    node: [18, 20, 22]
    os: [ubuntu-latest, windows-latest]
  fail-fast: false

#### 5. Reusable Workflows

All reusable workflows must be in root .github/workflows/ (no subdirectories).

# Caller
jobs:
  tests:
    uses: ./.github/workflows/reusable-test.yml
    with:
      package: server
    secrets:
      DATABASE_URL: ${{ secrets.DATABASE_URL }}

#### 6. Conditional Steps

- if: github.ref == 'refs/heads/main'
  run: pnpm deploy

- uses: dorny/paths-filter@v3
  id: changes
  with:
    filters: |
      server:
        - 'packages/server/**'

Debugging Pipeline Failures

Step 1: Get the failure

gh run list --branch main --limit 10
gh run view <run-id> --log-failed

Step 2: Classify

• SETUP: actions/checkout, setup-node, cache restore failed
• BUILD: compile error, missing dependency
• TEST: assertion failure, timeout
• DOCKER: image build, health check
• ENV: missing secret, wrong port
• PERMISSION: token scope, branch protection

Step 3: Common fixes

Optimization Patterns

#### Parallel Jobs

# Instead of: lint -> test -> build -> deploy (slow)
# Run in parallel:
# lint  -\
# test  ---> build -> deploy (fast)
# type  -/

#### Selective CI

- uses: dorny/paths-filter@v3
  id: filter
  with:
    filters: |
      e2e-needed:
        - 'packages/web-agent/**'

e2e:
  if: needs.filter.outputs.e2e-needed == 'true'

#### Artifact Sharing

- uses: actions/upload-artifact@v4
  with:
    name: build-output
    path: .next/
    retention-days: 1

Output Format

When creating or modifying a workflow, always provide:

1. Structure explanation -- what jobs exist, their order, why
2. Gotchas -- caching keys, port numbers, required secrets
3. Required secrets/vars list
4. Test command -- how to trigger and verify

Reference Files

Load these on demand for extended guidance:

Quality Gates

• Workflow YAML must be syntactically valid
• All required secrets must be documented
• Jobs must use proper dependency chains (needs)
• Caching must be configured for dependencies
• Concurrency must be set to prevent duplicate runs
• Changes must be explained with reasoning
• Test command must be provided for verification