uar-config-wizard
v1.0.0Configuration assistant for Universal Agent Runtime: guided wizard, advisor, validator, migration assistant, model selection advisor (models.dev + candle-vllm + turboquant), and full stack bundle generator
By Prometheus AGSLicense: MIT0 GitHub starsUpdated yesterday
Directory evidence
- Runtimes
- Claude Code
- Parsed components
- 8 skill or MCP entries
- Source updated
- Oct 4, 2026
- Manifest status
- Canonical path parsed
The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →
Install uar-config-wizard for Claude Code
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install uar-config-wizard@agent-plugin-marketplacePaste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.
The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.
Get the source manually
git clone https://github.com/Prometheus-AGS/universal-agent-runtimeClone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is .agent/skills/uar-config-wizard/.
Plugin files
├── .claude-plugin/plugin.json├── skills/uar-config/SKILL.md├── SKILL.md├── skills/uar-k8s-config/SKILL.md├── skills/uar-migrate/SKILL.md├── skills/uar-model-select/SKILL.md├── skills/uar-stack/SKILL.md├── skills/uar-validate/SKILL.md└── skills/uar-wizard/SKILL.md
Included Skills8
Main entry point for the UAR Configuration Wizard. Auto-detects intent and routes to the appropriate mode: wizard (first-time setup), validate (check existing config), migrate (upgrade legacy env vars), model-select (choose a model), k8s (generate Kubernetes manifests), or stack (full bundle).
Configuration assistant for Universal Agent Runtime (UAR). Covers all modes: interactive wizard (guided Q&A → tailored config.yaml + .env), advisor (explain options, generate annotated templates), validator (check existing config.yaml / .env for errors), migration assistant (upgrade LLM_* / legacy env vars to UAR_* prefix), Kubernetes config generation, model selection advisor (hardware profiling + live models.dev search via Tavily + candle-vllm catalog + turboquant KV-cache compression sizing), and full-stack bundle generator (UAR config.yaml + candle-vllm models.yaml + .env + quickstart.sh + optional K8s manifests). Generates immediately-runnable configs for every supported deployment scenario including local candle-vllm inference.
Generates Kubernetes manifests for UAR deployment: a Secret containing sensitive fields (jwt_secret, database_url, API keys) and a ConfigMap containing non-sensitive settings. Optionally also generates a candle-vllm Deployment manifest when used with the candle-vllm stack pattern.
Migrates legacy LLM_* and bare environment variable names to the UAR_* prefix convention. Scans .env files, detects deprecated variables, fixes provider/model format issues (e.g., "gpt-4o" → "openai/gpt-4o"), and produces an updated .env with a clear before/after diff. Backward compatible — both old and new vars work, UAR_* takes precedence.
AI model selection advisor for UAR + candle-vllm deployments. Five-phase guided flow: hardware profiling (GPU/VRAM/RAM/platform), use-case requirements (context, tools, vision, reasoning), live Tavily web research on models.dev + HuggingFace + liter-llm docs, VRAM-fit scoring + ranking, and TurboQuant KV-cache compression configuration. Produces a recommended model list with rationale, a candle-vllm models.yaml entry, and UAR provider config snippet.
Full-stack bundle generator for UAR + candle-vllm. Runs the guided wizard and model-select advisor in sequence, then emits a complete, immediately-runnable bundle: UAR config.yaml, candle-vllm models.yaml with TurboQuant kvcache_compression, .env with all secrets, quickstart.sh CLI script, and optional Kubernetes manifests. One command to configure the entire local LLM inference + agent runtime stack.
Validates an existing UAR config.yaml and/or .env file. Checks YAML syntax, required fields, type correctness, deprecated keys, security issues (default JWT secret detection), and candle-vllm integration correctness. Reports errors (❌), warnings (⚠️), and info (ℹ️) with exact fix instructions.
Guided first-time setup wizard for UAR. Asks targeted questions about deployment scenario, LLM provider, database backend, security, and optional features. Produces a tailored config.yaml and .env. If candle-vllm is chosen as the provider, automatically routes through model selection and TurboQuant configuration before generating the final files.
Plugin manifests1
{
"name": "uar-config-wizard",
"description": "Configuration assistant for Universal Agent Runtime: guided wizard, advisor, validator, migration assistant, model selection advisor (models.dev + candle-vllm + turboquant), and full stack bundle generator",
"version": "1.0.0",
"author": {
"name": "Prometheus AGS",
"url": "https://github.com/Prometheus-AGS"
},
"homepage": "https://github.com/Prometheus-AGS/universal-agent-runtime",
"repository": "https://github.com/Prometheus-AGS/universal-agent-runtime",
"license": "MIT",
"skills": [
"./skills/uar-config",
"./skills/uar-wizard",
"./skills/uar-validate",
"./skills/uar-migrate",
"./skills/uar-k8s-config",
"./skills/uar-model-select",
"./skills/uar-stack"
],
"agents": "./agents",
"hooks": "./hooks/hooks.json"
}For maintainers
If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.
[uar-config-wizard on Agent Plugins Marketplace](https://pluginsmp.com/plugins/uar-config-wizard)