HomeUse CasesSafe Ship Pipeline
New

Safe Ship Pipeline

Ship a change end to end without breaking production. This workflow chains three new plugins to make the risky database change safe, pre-flight the deploy manifests, prepare a clean commit and PR with the right version bump, and triage CI the moment it goes red.

3 Plugins
4 Steps
01

Make the migration safe

SSsql-safety-net

Before the schema change ships, check it for table-locking and blocking operations. SQL Safety Net rewrites unsafe migrations into zero-downtime expand-contract steps and emits a paired rollback - no database connection required.

$/migration-safetymigrations/0042_add_status.sql
Hazard tableSafe rewriteRollback migration
02

Pre-flight the deploy

CMci-incident-medic

Validate your Kubernetes manifests before they hit the cluster. The k8s-preflight check catches the silent failures - string-vs-int ports, missing resource limits, absent probes, and no securityContext - and returns corrected YAML.

$/k8s-preflightk8s/
PASS/WARN/FAIL gateCorrected manifestsProbe & limit fixes
03

Prepare the commit and PR

RCrelease-conductor

Turn your diff into a clean Conventional Commit, a recommended semver bump, and a complete PR description. Release Conductor dispatches the semver-advisor to reason about behavioral breaking changes, then drafts everything for you to approve.

$/shipmain
Conventional commitSemver recommendationPR description
04

Triage CI if it goes red

CMci-incident-medic

If the pipeline fails, don't spelunk through logs. The ci-failure-triager classifies the failure against a taxonomy of GitHub Actions failure modes - permissions, secrets, matrix, cache, triggers - and proposes the exact YAML fix.

$/gha-triage.github/workflows/ci.yml
Root causeEvidenceCorrected workflow

Getting the most out of this workflow

Run /migration-safety before anything else. Catching a table-locking ALTER before it reaches staging is the difference between a deploy and an incident.

Let /ship recommend the version bump. The semver-advisor reasons about behavioral breaking changes, not just syntax - it catches the ones humans get wrong about 40% of the time.

Keep the CI & Incident Medic hook enabled. When a build command fails, it offers /gha-triage automatically so you go from red to root cause without digging through logs.

Plugins in this workflow

SS
sql-safety-net
Step 1 - Safe migration
CM
ci-incident-medic
Steps 2 & 4 - Pre-flight & triage
RC
release-conductor
Step 3 - Commit, PR & version
Install all at once
/plugin install sql-safety-net
/plugin install ci-incident-medic
/plugin install release-conductor

Prerequisites

Claude Code installed and running
ClaudeRegistry marketplace added
~
A git repository with a change ready to ship
i
Works with any stack - the plugins detect your framework, dialect, and CI provider

Tags

releaseci-cddatabasedevopssafety

Have a workflow to share?

Submit your own use case and help other developers discover powerful plugin combinations.

Submit a Workflow →