Quick pick:

GitHub repo? → GitHub Actions
GitLab repo? → GitLab CI
Need extreme parallelism? → CircleCI
Azure/Microsoft shop? → Azure DevOps
Using Jira/Confluence? → Bitbucket Pipelines

Core Workflows

1. Pipeline Creation

Requirements → Platform Selection → Structure → Implementation → Validation

Steps:

Identify triggers (push, PR, schedule, manual)
Define stages (build, test, deploy)
Map environments (dev, staging, prod)
Configure secrets and variables
Set up caching strategy
Implement deployment gates

Minimal viable pipeline:

# Every pipeline needs these elements
triggers:        # When to run
stages:          # What to run (in order)
  - build        # Compile/bundle
  - test         # Validate
  - deploy       # Ship (optional)
caching:         # Speed optimization
environment:     # Secrets/variables

2. Pipeline Review

Review checklist (in order of priority):

Security
- Secrets not hardcoded
- Minimal permissions (least privilege)
- Dependencies pinned (no @latest)
- Untrusted input sanitized
Reliability
- Idempotent steps
- Explicit failure handling
- Timeouts configured
- Retry logic for flaky steps
Efficiency
- Caching implemented
- Parallelization where possible
- Conditional execution (skip unchanged)
- Resource sizing appropriate
Maintainability
- DRY (reusable workflows/templates)
- Clear naming conventions
- Comments on non-obvious logic

3. Pipeline Debugging

Diagnostic framework:

Symptom → Category → Platform-specific fix

Common failure categories:

Symptom	Likely Cause	First Check
"Command not found"	Missing dependency	Environment setup
"Permission denied"	File/secret access	Permissions config
"Connection refused"	Service not ready	Health checks, wait
"Out of memory"	Resource limits	Runner sizing
"Timeout exceeded"	Slow step or deadlock	Step isolation
"Exit code 1"	Script failure	Run locally first

Debug approach:

Read the full error (scroll up!)
Identify which step failed
Check if it works locally
Add debug output (set -x, verbose flags)
Check platform-specific logs

4. Deployment Strategy Selection

                     Risk Tolerance
                          ↑
                    Low   │   High
                          │
        ┌─────────────────┼─────────────────┐
        │                 │                 │
 Slow   │   Blue-Green    │    Rolling      │  Fast
        │   (safest)      │   (balanced)    │
        │                 │                 │
        ├─────────────────┼─────────────────┤
        │                 │                 │
        │    Canary       │   Direct        │
        │   (measured)    │   (fastest)     │
        │                 │                 │
        └─────────────────┴─────────────────┘
                     Rollback Speed →

Decision guide:

Strategy	Use When	Avoid When
Blue-Green	Zero downtime critical, simple rollback needed	Limited infrastructure budget
Canary	Need to validate with real traffic, gradual rollout	Urgent hotfix, no traffic splitting
Rolling	Large clusters, cost-conscious	Stateful apps, breaking changes
Direct	Dev/test environments, low risk	Production, user-facing services

5. Build Optimization

Optimization priority:

Caching (highest impact)
Parallelization
Conditional execution
Resource sizing
Image optimization

Caching hierarchy:

Dependencies (npm, pip) → Build artifacts → Docker layers → Test fixtures
       ↑                        ↑                ↑              ↑
   Always cache          If build slow     If using Docker   If tests slow

Cache key strategies:

# Dependencies - hash lockfile
key: deps-${{ hashFiles('package-lock.json') }}

# Build artifacts - hash source
key: build-${{ hashFiles('src/**') }}

# Combined - hash both
key: ${{ runner.os }}-${{ hashFiles('**/lockfiles') }}

Common Patterns

Secret Management

# ✅ Good: Environment variables
env:
  API_KEY: ${{ secrets.API_KEY }}

# ❌ Bad: Inline secrets
run: curl -H "Authorization: hardcoded-secret"

Secret rotation checklist:

Secrets stored in platform secret store
Secrets scoped to minimum required (repo/org/env)
Rotation procedure documented
No secrets in logs (mask sensitive output)

Environment Progression

┌─────────┐    ┌─────────┐    ┌─────────┐    ┌─────────┐
│   Dev   │ →  │Staging  │ →  │   QA    │ →  │  Prod   │
└─────────┘    └─────────┘    └─────────┘    └─────────┘
  Auto          Auto          Manual          Manual
  deploy        deploy        approval        approval

Artifact Management

# Upload
- name: Upload build
  uses: actions/upload-artifact@v4
  with:
    name: dist
    path: dist/
    retention-days: 7

# Download (later job)
- name: Download build
  uses: actions/download-artifact@v4
  with:
    name: dist

Matrix Builds

# Test across multiple versions
strategy:
  matrix:
    node: [18, 20, 22]
    os: [ubuntu-latest, macos-latest]
  fail-fast: false  # Continue other jobs if one fails

Anti-Patterns

❌ Secrets in Logs

# BAD
run: echo "Deploying with key $API_KEY"

# GOOD
run: echo "Deploying..." && deploy --key "$API_KEY"

❌ No Caching

# BAD: Downloads deps every run
- run: npm install

# GOOD: Cache node_modules
- uses: actions/cache@v4
  with:
    path: node_modules
    key: ${{ hashFiles('package-lock.json') }}
- run: npm ci

❌ Mutable Tags

# BAD: Could change anytime
uses: actions/checkout@master

# GOOD: Pinned version
uses: actions/checkout@v4

❌ Overly Broad Triggers

# BAD: Runs on every push to every branch
on: push

# GOOD: Specific triggers
on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

❌ Sequential When Parallel Works

# BAD: 15 min total
jobs:
  lint:
    ...
  test:
    needs: lint  # Waits for lint
    ...

# GOOD: 10 min total (parallel)
jobs:
  lint:
    ...
  test:
    ...  # Runs alongside lint

Troubleshooting Quick Reference

Error Pattern	Platform	Solution
`EACCES: permission denied`	All	`chmod +x script.sh` or check file ownership
`rate limit exceeded`	GitHub	Use `GITHUB_TOKEN`, implement backoff
`no space left on device`	All	Clean workspace, use smaller runner
`SSL certificate problem`	All	Update CA certs, check proxy settings
`container not found`	Docker-based	Check registry auth, image exists
`workflow not triggered`	GitHub	Check trigger conditions, branch patterns
`job timeout`	All	Increase timeout, optimize slow steps

Integration Points

With version control:

Branch protection rules
Required status checks
Auto-merge on success

With deployment targets:

Cloud providers (AWS, GCP, Azure)
Container registries (Docker Hub, ECR, GCR)
Kubernetes clusters
Serverless (Lambda, Cloud Functions)

With monitoring:

Build metrics (duration, success rate)
Deployment tracking
Error alerting (Slack, PagerDuty)

Skill Usage

Creating a pipeline:

Identify platform from repo hosting
Read platform-specific reference
Follow creation workflow above
Validate with platform's linter

Debugging a failure:

Categorize the error (table above)
Read debugging-guide.md for detailed steps
Check platform-specific quirks

Implementing deployments:

Choose strategy using decision guide
Read deployment-strategies.md
Implement with platform-specific syntax

Security audit:

Use review checklist above
Read security-checklist.md for comprehensive review
Generate ohno tasks for findings

References:

references/github-actions.md — GitHub Actions syntax, patterns, marketplace actions
references/gitlab-ci.md — GitLab CI syntax, Auto DevOps, runners
references/circleci.md — CircleCI orbs, parallelism, caching
references/azure-devops.md — Azure Pipelines, templates, stages
references/bitbucket-pipelines.md — Bitbucket Pipelines, pipes, deployments
references/deployment-strategies.md — Blue-green, canary, rolling implementations
references/debugging-guide.md — Platform-specific debugging techniques
references/security-checklist.md — OWASP CI/CD security, supply chain

Platform	Best For	Key Strength
GitHub Actions	GitHub-hosted repos, open source	Native GitHub integration, marketplace
GitLab CI	GitLab repos, self-hosted	Built-in registry, Auto DevOps
CircleCI	Complex workflows, speed	Parallelism, orbs ecosystem
Azure DevOps	Microsoft/enterprise, multi-repo	Azure integration, YAML templates
Bitbucket	Atlassian stack, Jira integration	Pipes marketplace, deployments

ci-cd-expert

CI/CD Expert

Platform Selection