local-llm-router — Claude Code skill: what it does, how to install
local-llm-router routes coding prompts to local language models inside air-gapped setups. Employs Serena MCP for codebase context without cloud access.
no license ★ 301 updated 2026-10-23
npx skills add hoodini/ai-agents-skills --skill local-llm-router -a claude-codeWhat local-llm-router does
This skill directs programming tasks to local runtimes such as Ollama, LM Studio, or Jan. Before passing prompts to any neural network, it invokes Serena MCP. The server inspects code structure via LSP primitives, extracting symbol definitions and references instead of dumping raw file trees.
The routine classifies incoming tasks into code generation, reasoning, analysis, or documentation. It scans local endpoints (11434, 1234, or 1337), selects the relevant model, and executes the request offline.
Without an active Serena MCP instance and at least one local server running, the router fails immediately. The entire workflow stays strictly on your machine.
What it helps with
- write code with AI assistance inside an air-gapped network
- dispatch requests between Ollama, LM Studio, and Jan
- gather semantic AST context through Serena MCP without string matching
Example prompts for the agent
- Route this refactoring request to local Ollama using Serena context
- Send the call hierarchy analysis to LM Studio
- Generate unit tests for this parser using the active Jan model
How to install local-llm-router
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill local-llm-router -a claude-code2Manually: put the files in place
~/.claude/skills/local-llm-router/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/local-llm-router/
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill local-llm-router -a codex2Manually: put the files in place
.agents/skills/local-llm-router/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill local-llm-router -a cursor2Manually: put the files in place
.cursor/skills/local-llm-router/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install hoodini/ai-agents-skills skills/local-llm-router2Manually: put the files in place
.github/skills/local-llm-router/SKILL.md1Skills CLI (Vercel)
npx skills add hoodini/ai-agents-skills --skill local-llm-router -a gemini-cli2Platform CLI
gemini skills install https://github.com/hoodini/ai-agents-skills --path skills/local-llm-routerManually: put the files in place
.devin/skills/local-llm-router/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about local-llm-router
How to install local-llm-router in Claude Code?
Run npx skills add hoodini/ai-agents-skills --skill local-llm-router -a claude-code in your terminal. The configuration is stored at ~/.claude/skills/local-llm-router/SKILL.md.
What does local-llm-router do?
It identifies reachable local model servers, gathers semantic code context using Serena MCP, and dispatches the task to the right offline engine.
What dependencies must run beforehand?
You need Serena MCP running alongside an active backend listening on http://localhost:11434, http://localhost:1234, or http://localhost:1337.