Implement the following plan:

Unified deploy commands

Goal

Replace the multi-step deployment process with bun run deploy:dev and bun run deploy:prod. Each command runs the full pipeline (migrate → build → push → restart) with quiet output on success and full output on failure.

New file: scripts/deploy.sh

Single shell script, takes dev or prod as first argument. Runs these steps in order:

1. Migrate database — port-forward to K8s postgres, run drizzle migrations against the appropriate database (dev or prod)
2. Build Docker images — docker build for web, api, mail with the appropriate tag (dev or latest+version)
3. Push Docker images — docker push all built images
4. Restart deployments — kubectl rollout restart for web and api, wait for rollout completion

Output strategy

Each step prints a single-line status while running (e.g., Migrating dev database...). On success, overwrite with a checkmark line. On failure, dump the captured output and exit.

Pattern per step:

step "Migrating database" migrate_db

Where step is a helper that:

• Prints "● Step name..."
• Runs the function, capturing stdout/stderr to a temp file
• On success: prints "✓ Step name"
• On failure: prints "✗ Step name", dumps captured output, exits 1

Key details

• Gets DB_PASSWORD from K8s secret, constructs DATABASE_URL for the target db
• Port-forward started/stopped automatically within the migrate step (background process with trap cleanup)
• Docker tag: dev for dev, latest+version for prod (version from apps/web/package.json)
• Checks Docker Hub login before push step
• Restarts web + api deployments (not mail — it rarely changes; can add --all flag later if needed)

Changes to package.json

Add two scripts, remove obsolete individual k8s scripts:

"deploy:dev": "./scripts/deploy.sh dev",
"deploy:prod": "./scripts/deploy.sh prod",
"restart:dev": "kubectl rollout restart deployment/dev-web deployment/dev-api -n usingandunderstandingai && kubectl rollout status deployment/dev-web deployment/dev-api -n usingandunderstandingai",
"restart:prod": "kubectl rollout restart deployment/web deployment/api -n usingandunderstandingai && kubectl rollout status deployment/web deployment/api -n usingandunderstandingai",

Remove these scripts (the underlying files stay for now, but they're no longer the primary interface):

• k8s:build
• k8s:push
• k8s:build-push

Keep these (used for initial cluster setup, not regular deploys):

• k8s:deploy
• k8s:teardown

Changes to CLAUDE.md

Update the deployment section to show the new commands as the primary interface:

bun run deploy:dev     # Full deploy to dev (migrate + build + push + restart)
bun run deploy:prod    # Full deploy to prod
bun run restart:dev    # Quick restart (no rebuild)
bun run restart:prod   # Quick restart (no rebuild)

Files changed

Verification

1. bun run deploy:dev — should migrate dev db, build dev images, push, restart dev deployments with minimal output
2. bun run deploy:prod — same for prod
3. bun run restart:dev / restart:prod — quick restart without rebuild
4. Intentionally break a step (e.g., kill port-forward) to verify failure output is shown

If you need specific details from before exiting plan mode (like exact code snippets, error messages, or content you generated), read the full transcript at: ~/.claude/projects/-Users-challen-www-usingandunderstanding-ai/06692091-d3c7-4fd9-993c-4d51d96caa44.jsonl

Your task

First, run git status, git diff, and git log --oneline -10 to understand the current state.

Then add and commit all modified files with a descriptive message. Break the changes into multiple commits if appropriate. If there are untracked files, ask if they should be added or added to the appropriate .gitignore file. The user may also want to take over at this point, so provide an abort option as well. Always check changes for unnecessary debugging statements or logging that might have been accidentally included. Prompt the user if you find any, and offer to remove each one by one. Do not use interactive git staging (git add -p) as it does not work.

OK. The deployment script should also offer the option of bumping the project versions, similar to what's done in ../cs124.org/k8s/deploy.ts. Can it do that?

Wait, what? Why is this a shell script? Implement it as a ts script please. Bun can run those directly.

This script does not need to be able to be run directly. It should only be run by using bun run deploy:dev.

Let's migrate the bin shell scripts to be .ts scripts that are called from package.json commands, accepting arguments as appropriate. There's a lot of duplication there.