ops-checklist

Run this before any production deployment to confirm the environment is healthy.

Steps

1. Check recent Sentry errors

Use the Sentry MCP to check for new errors in the last 24 hours:

• Query the casedive project for unresolved issues with firstSeen:>-24h
• Flag any issues with times_seen > 10 or severity fatal/error
• If critical errors exist, stop and surface them — do not proceed until acknowledged

2. Verify env vars are present on Vercel

vercel env ls

Confirm these are present (production + preview):

• ANTHROPIC_API_KEY
• UPSTASH_REDIS_REST_URL
• UPSTASH_REDIS_REST_TOKEN
• CANLII_API_KEY (optional but flag if missing)
• SENTRY_DSN

If any required var is absent, stop and raise it.

3. Upstash quota reminder

Manually check Upstash console for:

• Daily command count approaching limit
• Any rate-limit errors in recent Redis logs

This step cannot be automated — note it and ask the user to confirm they've checked it.

4. Build verification

npm run build

Must exit 0. If it fails, do not deploy.

5. Security scan

npm run security:scan

Must pass with no new findings beyond known false positives (see memory: AgentShield chmod false positive).

6. Final checklist

Before signalling ready-to-deploy:

• No critical Sentry errors in last 24h (or acknowledged)
• All required env vars present on Vercel
• Upstash quota confirmed healthy (manual)
• npm run build passes
• npm run security:scan passes
• No uncommitted secrets in working tree (git status)

If all items pass, output: Ops checklist complete — safe to deploy.
If any item fails, output the specific blocker and stop.