deploy-flow

Governs production rollouts on GitHub covering pipelines, image tagging, migrations, and cleanup.

1|Updated Jan 14, 2024
One-click install
npx skills add https://github.com/Eyhenij/rt-tools --skill deploy-flow-eyhenij
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deploy-flow
Source: https://github.com/Eyhenij/rt-tools/tree/main/.claude/skills/deploy-flow
Command: npx skills add https://github.com/Eyhenij/rt-tools --skill deploy-flow-eyhenij

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams shipping to production from a GitHub repository often lack explicit rules for how rollouts actually work: when a merge does or does not deploy, how images are tagged, how migrations are verified, and how failures are detected. This Skill encodes those delivery rules so an AI assistant applies them consistently whenever an edit goes to production. ## Core Features & Use Cases - Rollout governance: Defines how merges, manual workflow_dispatch runs, image sha tagging, and rollback depth (last three shas) are handled in a GitHub Actions pipeline. - Migration safety: Requires running the migration chain from an empty storage before merge and measuring schema drift on a shadow database rather than the developer's working one. - Divergence auditing: Detects when production lags the main branch via a work queue audit that reads the last successful rollout, catching failed or never-started deployments. - Use Case: Before merging a PR that changes the deploy workflow or database schema, load this rule so the assistant enforces pre-merge pipeline runs, shadow-storage migration checks, and sha-based image rollout instead of the 'latest' tag. ## Quick Start Load the deploy-flow rule and review my pull request that changes the deploy workflow and adds a database migration before I merge it into main.

Frequently Asked Questions about deploy-flow

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy to production from a GitHub repository safely?

Deploy by rolling out images tagged with the commit sha rather than the 'latest' tag, and run the production stack locally before merging. Verify what is actually deployed by querying the last successful workflow run with gh run list and comparing its sha against the main branch.

How to run database migrations before merging a pull request?

Run the full migration chain from an empty storage before the merge, since apply order is lexicographic by directory name and timestamps are set at creation. Measure schema drift on a shadow database, not the developer's working one, so unfinished branches do not block the check.

Why should Docker images be deployed by commit sha instead of latest?

The 'latest' tag in the registry lags behind the main branch, so production can silently fall back to a previous version while still answering requests. Rolling out by commit sha guarantees the deployed image matches the intended commit and keeps a defined rollback depth.

Does a merge into main automatically deploy to production?

Not necessarily. In this tree the deploy workflow has only a workflow_dispatch trigger, so a merge rolls out nothing until someone starts the workflow manually. The rule requires each tree to name what starts its rollout rather than assuming merge-based deployment.

How do I detect when production lags behind the main branch?

Use a work queue audit that queries the last successful rollout workflow run and counts commits between its sha and the main branch tip. A main-branch CI run alone says nothing about production when rollouts are started manually.