kelly-llm-gateway

Monitor shared LLM gateway traffic, costs, and canary rollouts via a local dashboard.

4|3|Updated Feb 25, 2026
One-click install
npx skills add https://github.com/mr-kelly/skills --skill kelly-llm-gateway
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kelly-llm-gateway
Source: https://github.com/mr-kelly/skills/tree/main/skills/kelly-llm-gateway
Command: npx skills add https://github.com/mr-kelly/skills --skill kelly-llm-gateway

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires hono, @hono/node-server, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This skill provides a centralized, brand-free dashboard for platform teams to monitor and govern shared LLM gateway traffic, costs, and canary rollouts without needing to modify live routing configurations.

Core Features & Use Cases

  • Cost & Performance Monitoring: Aggregates daily spend trends, call volumes, and error rates across multiple services and models.
  • Canary Rollout Management: Provides a visual status board to track canary percentages and perform promote, rollback, or hold actions.
  • Anomaly Detection: Automatically identifies cost and error-rate spikes using deterministic rolling baselines.
  • Use Case: A platform engineer can use this tool to review a canary rollout for a new internal model, verify that error rates remain within thresholds, and promote the model to 100% traffic without touching the production gateway config.

Quick Start

Use the kelly-llm-gateway skill to open the local dashboard and review the current gateway cost and rollout status.

Frequently Asked Questions about kelly-llm-gateway

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I monitor LLM gateway costs and error rates locally?

You can monitor LLM gateway costs and error rates using a local dashboard that aggregates daily spend trends, call volumes, and error rates across multiple services and models.

What is the best way to manage canary rollouts for internal LLM models?

Managing canary rollouts for internal LLM models involves using a visual status board to track rollout percentages and perform promote, rollback, or hold actions without touching live routing configurations.

How does deterministic anomaly detection work for LLM traffic?

Deterministic anomaly detection for LLM traffic works by automatically identifying cost and error-rate spikes using rolling baselines, allowing platform teams to acknowledge anomalies directly.

Can I promote a canary model to 100% traffic without modifying production gateway config?

Yes, you can promote a canary model to 100% traffic without modifying production gateway config by using a local file-backed interface that generates handoff files for human-led configuration changes.

Does the Hono framework support local LLM gateway governance dashboards?

Yes, the Hono framework supports local LLM gateway governance dashboards by providing a local, brand-free interface for platform teams to monitor shared traffic and costs.

What are the limitations of using a local file-backed interface for LLM governance?

The limitation of a local file-backed interface for LLM governance is that it does not modify live routing configurations directly, but instead generates handoff files requiring human-led configuration changes.