create-a-deployment-pipelinelisted

# create-a-deployment-pipeline Turn deployment into an inspectable system. The pipeline should make the safe path easier than the improvised path and leave enough evidence to answer what changed, who approved it, which artifact ran, and how it was checked. ## When to use Use this skill when a team deploys manually, environments drift, releases cannot be reproduced, or production changes need review and recovery. Do not add stages that have no failure they can detect or decision they can support. Never print secrets into logs or pass unreviewed production credentials to forked code. Separate build permissions from deployment permissions and keep an emergency path narrow and audited. ## Preconditions - Map the current path from commit to production, including manual commands and hidden credentials. - Name environments, deployment owner, approval authority, service health checks, and rollback owner. - Define the source revision, dependency lock, build environment, and artifact retention policy. - Identify database, configuration, and infrastructure changes that cannot be rolled back like code. ## Procedure 1. Choose one version identifier that connects commit, immutable artifact, release record, and runtime. 2. Make CI build and test the artifact in a clean environment with pinned dependencies. 3. Store the artifact once. Promote the same verified bytes across environments instead of rebuilding. 4. Add gates for unit tests, static checks, secret scanning, artifact integr