CI/CD
Apply CI/CD to guide decision-making, structure analysis, and improve outcomes in relevant contexts.
SKILL.mdcomposable-skills/frameworks/cicd/SKILL.md
GitHubCI/CD
Purpose
Use continuous integration and continuous delivery to make releasing software routine, low-risk and fast. Continuous integration means everyone merges small changes to a shared mainline frequently, with an automated build and test run each time. Continuous delivery means every change that passes the pipeline is releasable, so deployment becomes a business decision rather than a technical event. Research by the DORA program has linked these practices to both speed and stability, showing that the two need not be traded off.
Frameworks & Standards
| Item | Value |
|---|---|
| Framework ID | cicd |
| Category | Engineering — Delivery Pipeline |
| Version | 2.0.0 |
| Key references | Fowler, "Continuous Integration" (2006) · Humble & Farley, Continuous Delivery (2010) · Forsgren, Humble & Kim, Accelerate (2018) |
| Delivery measures | The DORA metrics. Not to be confused with DORA, the EU Digital Operational Resilience Act, which has its own definition. |
The Practices
- Mainline development: small, frequent merges, ideally daily.
- Automated build and test on every change.
- A deployment pipeline: stages from commit to production, each giving more confidence.
- Build once, deploy the same artifact to every environment.
- Deployment automation and safe release techniques: feature flags, canary and blue-green releases.
- Fast feedback: a broken build is fixed or reverted immediately.
The DORA software delivery metrics
DORA currently defines five, in two groups (dora.dev):
- Throughput: change lead time; deployment frequency; failed deployment recovery time.
- Instability: change fail rate; deployment rework rate. Earlier DORA reports and Accelerate (2018) used four metrics, with "time to restore service" in place of failed deployment recovery time and no rework rate.
Core Principles
- Small batches reduce risk. Smaller changes are easier to test, review and roll back.
- Keep the mainline releasable at all times.
- Automate everything repeatable, from build to deployment.
- Fix a broken pipeline first. A red build blocks everyone.
- Speed and stability go together when the pipeline is well built.
Applications & Use Cases
| Use Case | Application | Practice |
|---|---|---|
| Painful, infrequent releases | Automate the pipeline and release smaller changes | Small batches |
| Long-lived branches | Move to trunk-based development | Mainline development |
| Environment drift | Build once, deploy the same artifact | Artifact promotion |
| Risky deployments | Canary or blue-green releases with feature flags | Safe release |
| Measuring delivery | Track the DORA software delivery metrics | Measurement |
| Quality gates | Tests, security and performance checks in the pipeline | Automated testing |
Usage Guidelines
- Run the full automated test suite on every merge to the mainline.
- Keep branches short-lived, merging at least daily.
- Promote one immutable artifact through all environments.
- Separate deployment from release with feature flags.
- Track the DORA metrics and improve the weakest.
Prompt Template
You are designing or reviewing a CI/CD pipeline (Humble & Farley 2010; DORA metrics).
CURRENT STATE: branch strategy · merge frequency · test automation · release process
PIPELINE STAGES (commit -> production):
ARTIFACT PROMOTION: built once?
RELEASE SAFETY: feature flags · canary · blue-green · rollback
DORA METRICS: change lead time · deployment frequency · failed deployment recovery time · change fail rate · deployment rework rate
BIGGEST CONSTRAINT and the next improvement:
Ethical Guidelines
- Protect pipeline secrets and signing keys; a compromised pipeline compromises every user.
- Do not bypass quality and security gates under deadline pressure.
Testing Strategy
- Layered tests: fast unit tests on commit, broader tests in later stages.
- Test the rollback path, not only the deployment path.
Success Metrics
- The DORA software delivery metrics (throughput and instability).
- Pipeline duration and reliability.
- Share of changes deployed without manual steps.
Collaboration Protocol
- Inbound from:
devops-automator·backend-architect·api-tester - Outbound to:
infrastructure-maintainerfor environments - Cross-framework: DevSecOps adds security to the pipeline; infrastructure as code provisions environments; Kanban shortens the flow before the pipeline.
Related Skills
composable-skills/frameworks/devsecops/SKILL.md— security in the pipelinecomposable-skills/frameworks/infrastructure-as-code/SKILL.md— reproducible environmentscomposable-skills/frameworks/kanban/SKILL.md— flow of work into the pipeline
Reference Materials
- Fowler, M. (2006). "Continuous Integration." martinfowler.com.
- Humble, J., & Farley, D. (2010). Continuous Delivery. Addison-Wesley.
- DORA. "DORA's software delivery performance metrics." dora.dev/guides/dora-metrics.
- Forsgren, N., Humble, J., & Kim, G. (2018). Accelerate: The Science of Lean Software and DevOps. IT Revolution Press.
Copyright (c) 2026 iSystematic Inc. Maxim product. BSL 1.1. Authored 2026-10-15, replacing a generated template stub.