ci-pipeline

name	ci-pipeline
description	Generates a CI/CD pipeline configuration for the project. Auto-detects the stack and creates GitHub Actions, GitLab CI, or other CI configs with lint, test, build, and deploy stages.
argument-hint	[github-actions, gitlab-ci, or 'auto']
allowed-tools	Read, Grep, Glob, Bash, Write

CI Pipeline Generator

Detect Project Stack

!ls package.json requirements.txt pyproject.toml go.mod Cargo.toml *.csproj *.sln Gemfile composer.json mix.exs Makefile Dockerfile 2>/dev/null !ls .github/workflows/ .gitlab-ci.yml Jenkinsfile .circleci/ .travis.yml bitbucket-pipelines.yml 2>/dev/null !cat package.json 2>/dev/null | grep -E '"(scripts|devDependencies)"' -A 10 | head -20

---

Pipeline Design

Step 1: Detect Requirements

From project files, determine:

• Language and version (Node 20, Python 3.12, Go 1.22, etc.)
• Package manager (npm, pnpm, yarn, pip, poetry, cargo, etc.)
• Test command
• Lint command
• Build command
• Required services (database, Redis, etc.)

Step 2: Generate Pipeline Stages

Stage 1: Install

• Cache dependencies for fast subsequent runs
• Use lock file hash as cache key

Stage 2: Lint

• Run linter (ESLint, ruff, golangci-lint, clippy, etc.)
• Run type checker if applicable (tsc, mypy, go vet)
• Run formatter check (prettier, black, gofmt)

Stage 3: Test

• Run unit tests with coverage
• Run integration tests (with service containers if needed)
• Upload coverage report

Stage 4: Build

• Production build
• Verify build artifacts are created

Stage 5: Security (optional)

• Dependency vulnerability scan
• Secret scanning
• SAST if available

Stage 6: Deploy (placeholder)

• Staging deployment (on push to main)
• Production deployment (on tag/release)
• Mark as manual/approval required

Step 3: Platform-Specific Output

GitHub Actions:

name: CI
on: [push, pull_request]
jobs:
  ci:
    runs-on: ubuntu-latest
    steps: ...

GitLab CI:

stages: [install, lint, test, build, deploy]

Step 4: Optimizations

• Dependency caching (npm cache, pip cache, cargo cache)
• Parallel test execution where possible
• Matrix builds for multiple versions (if needed)
• Fail-fast on lint errors (don't waste time on tests)
• Artifact upload for build output

Rules

• Always include a lint stage — catch issues early
• Always include caching — speed up repeat runs
• Keep the pipeline under 10 minutes for PRs
• Don't hardcode versions — use variables or matrix
• Include both push and PR triggers