Modernize: Bolt 4, Node 24, error handling + GHCR/Coolify deploy

[ddbruce]

Summary

Behavior-preserving modernization of watchdog. Same behavior: a message from a non-admin/non-moderator in a monitored channel is deleted and the user gets the same ephemeral warning; thread replies and message_deleted events are ignored. Verified: 8 unit tests pass; app boots on Bolt 4 with /health → 200 and unsigned /slack/events → 401; amd64 image builds with 0 npm vulnerabilities.

Changes

• SDK/runtime: @slack/bolt 3.6 → 4.4, dotenv 10 → 17; Dockerfile node:16-alpine (EOL 2023) → pinned node:24-alpine, npm ci, non-root, HEALTHCHECK.
• Health endpoint: custom HTTPReceiver with a GET /health route. The Slack events path (/slack/events) and its signing-secret verification are unchanged.
• Correctness: app.message handler is now async and awaits processMessage; added an app.error handler; chat.delete is now awaited inside try/catch (was fire-and-forget); MONITORED_CHANNELS parsing is guarded so a missing env var no longer throws at startup.
• Testability: pure decision logic extracted into src/utilities/moderation.js (no Slack import) with Jest tests. The admin-list short-circuit in isNotModerator (mod group only fetched if the user isn't an admin) is preserved exactly.
• Tooling/CI: ESLint 7 (EOL) → ESLint 9 flat config; dead Drone CI → GitHub Actions (lint + test + amd64 build).
• Deploy: scripts/deploy.sh + config.example.sh (amd64 → GHCR → Coolify POST /api/v1/deploy?uuid=); secrets in gitignored scripts/config.sh.

Notes

• No env-var changes — same variables as before.
• On Bolt 4, an invalid/expired bot token now makes the app fail fast at startup (Bolt runs auth.test on start). With valid creds this is a no-op; worth knowing for crash-loop debugging.

Recommended follow-up (NOT done — would change behavior)

• Cache usergroup membership with a short TTL to cut two Slack API calls per message. Skipped because it delays when admin/mod role changes take effect; happy to add if you want it.