You are doing an emergency production hotfix. Move fast but don't break more things.

Step 1: Diagnose (2 min max)

First understand the problem from $ARGUMENTS. Then check logs:

API Hub (Cloud Run):

gcloud run services logs read api-hub --limit=50 --region=us-central1 2>/dev/null | tail -30

Site (Vercel):

# Check if site is up
curl -s -o /dev/null -w "%{http_code}" https://skillboss.co

Quick health check:

curl -s https://api.heybossai.com/health 2>/dev/null || echo "API Hub unreachable"

Identify the root cause before touching code. State it clearly: "Root cause: ..."

Step 2: Fix

• Make the minimal change that fixes the issue — no refactoring, no cleanup
• Touch as few files as possible
• If unsure between two approaches, pick the safer/more reversible one

Step 3: Verify locally

api-hub:

python3 -m py_compile <changed files>

onepersoncompany:

cd /Users/xiaoyinqu/skillboss && pnpm typecheck 2>&1 | tail -5

Step 4: Deploy

api-hub (Google Cloud Run):

cd /Users/xiaoyinqu/heyboss/api-hub
git add <changed files>
git commit -m "hotfix: $ARGUMENTS"
git push origin main
gcloud run deploy api-hub --source . --region us-central1 --quiet 2>&1 | tail -10

onepersoncompany (Vercel — auto-deploys on push):

cd /Users/xiaoyinqu/skillboss
git add <changed files>
git commit -m "hotfix: $ARGUMENTS"
git push origin main

Step 5: Verify production

Wait 30 seconds, then:

curl -s https://api.heybossai.com/health
or test the specific broken endpoint

Report: "✅ Fixed: [what was broken → what you did → confirmed working]"

---

Hotfix rules:

• Don't create a branch — push direct to main (this is an emergency)
• Don't wait for CI — verify manually
• Write a follow-up ticket/note for proper cleanup later
• If it's not fixed in 10 min, roll back: git revert HEAD && git push origin main