chore(plugins): remove non-functional Zomato Omi app

[kodjima33]

What

Removes the plugins/omi-zomato-app/ plugin entirely (FastAPI service + Railway deployment wiring: main.py, mcp_proxy.py, db.py, models.py, railway.toml, Procfile, requirements.txt, .gitignore).

Why

Per issue #8905, the Zomato app is non-functional and has never had a successful production deployment:

Total deployments checked: 1,885
Successful deployments found: 0
Earliest recorded deployment (4076445966) failed at 2026-03-15T20:15:16Z
Latest deployment (5295112471) failed at 2026-07-03T05:22:12Z

The issue explicitly frames this as a cleanup/removal, not a repair: "If the integration is still desired, it should probably be rebuilt as a new app with a successful deployment path."

What I verified

• All zomato references in the repo were self-contained within plugins/omi-zomato-app/. After removal, git grep -il zomato returns no matches (exit 1).
• No plugin registry, app-store listing, or shared config outside the plugin directory referenced the app — the plugin's Railway/deployment wiring (railway.toml, Procfile) lived inside the removed directory.
• Pure deletion of dead code; no runtime paths in other services touch it.

fixes #8905

Auto-generated from issue feedback by the mini issues-improver. Review before merge.

[cubic-dev-ai]

No issues found across 8 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}

[Git-on-my-level]

Maintainer review (comment-only, not a formal approval):

Thanks for cleaning this up. I reviewed the PR statically against issue #8905. The diff is a pure deletion of plugins/omi-zomato-app/ including its FastAPI service, MCP/OAuth proxy, Redis/file token storage, Railway config, Procfile, requirements, and local gitignore. I also checked the PR head for remaining zomato / omi-zomato-app references and found none.

Positive signal: this looks like a coherent cleanup of a non-functional, self-contained app, and the current checks are green.

I’m leaving this as comment-only rather than approving because it removes an integration and dependency/deployment surface; a human maintainer should confirm that deleting the app is the intended product direction before merge. If that intent is confirmed, I do not see code-level blockers in this PR.