EnigmAgent Plugin for LM Studio

TypeScript plugin that exposes three LM Studio tools:

| Tool | Description | |------|-------------| | enigmagent_resolve | Resolve {{PLACEHOLDER}} tokens in any text string | | enigmagent_list | List secret names available in the vault (never values) | | enigmagent_status | Check vault running/unlocked state |

Prerequisites

1. EnigmAgent vault running: enigmagent-mcp --mode rest --port 3737
2. Node.js 18+ / npm

Install & Build

cd platforms/integrations/lm-studio/lmstudio-plugin
npm install
npm run build

Register in LM Studio

LM Studio does not yet have a plugin registry UI. Register via the SDK:

import { LMStudioClient } from "@lmstudio/sdk";
import { tools } from "./dist/index.js";

const client = new LMStudioClient();
// Load a model and attach the tools
const model = await client.llm.load("lmstudio-community/Meta-Llama-3-8B-Instruct-GGUF");
const response = await model.respond(
  [{ role: "user", content: "List my available secrets: call enigmagent_list" }],
  { tools }
);
console.log(response.content);

Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | ENIGMAGENT_HOST | 127.0.0.1 | Vault REST API host | | ENIGMAGENT_PORT | 3737 | Vault REST API port | | ENIGMAGENT_TOKEN | (empty) | Bearer token (optional) |

Example

import { tools } from "enigmagent-lmstudio-plugin";

// The model can call enigmagent_resolve automatically:
// User: "Send a request using {{OPENAI_KEY}}"
// → model calls enigmagent_resolve({ text: "Send a request using {{OPENAI_KEY}}" })
// → vault returns real value
// → model uses resolved text