# Authenticated LLM Agent

The Authenticated LLM Agent MCP server checks JWT credentials before sending a request to Gemini through LangChain and Vertex. Its separate self-test tool returns a sample result and trust details.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · 2349 installs
- Author: a2a
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://smithery.ai/server/a2a/authenticated-llm-agent
- Updated: 2026-05-31
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/mcp/authenticated-llm-agent

## What it does

The `invoke` tool checks a password with bcrypt and validates JWT credentials before requesting a Gemini completion through LangChain on Vertex. Requests without authentication are rejected. Access checks are part of the call flow.

Use `selftest` to run the agent on a fixed input. It returns the actual result, an assertion grade, and trust information: a signed Agent Card, JWKS, a governance and licence chain, and a safety posture. Each claim has a proof link. The test is free and requires no payment.

## Requirements
- Remote server: `https://authenticated-llm-agent--a2a.run.tools`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http authenticated-llm-agent https://authenticated-llm-agent--a2a.run.tools
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add authenticated-llm-agent --url https://authenticated-llm-agent--a2a.run.tools
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "authenticated-llm-agent": {
            "url": "https://authenticated-llm-agent--a2a.run.tools"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "authenticated-llm-agent": {
            "type": "http",
            "url": "https://authenticated-llm-agent--a2a.run.tools"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http authenticated-llm-agent https://authenticated-llm-agent--a2a.run.tools
```

## When to use it
- gate LLM requests behind JWT authentication
- reject calls that lack valid authentication
- run the agent against a fixed test input
- inspect trust claims alongside proof links

## Example prompts
- Use `invoke` to send a Gemini request after JWT authentication.
- Run `selftest` with its fixed input and show me the result.
- Show the assertion grade and proof links returned by `selftest`.

## Similar
- [Secured LangChain OpenAI Agent](https://affpapa.org/en/nevibecode/mcp/secured-langchain-openai-agent.md) — MCP server · 2894 installs
- [LLM Observability & Orchestration Agent](https://affpapa.org/en/nevibecode/mcp/llm-observability-orchestration-agent.md) — MCP server · 2409 installs
- [llm-app-development](https://affpapa.org/en/nevibecode/skills/llm-app-development.md) — skill · 139 installs
- [llm-application-dev](https://affpapa.org/en/nevibecode/plugins/seth-hobson-llm-application-dev.md) — plugin · ★ 126 repo stars
- [llm-models](https://affpapa.org/en/nevibecode/skills/llm-models.md) — skill · 238 installs
- [llm-cli](https://affpapa.org/en/nevibecode/skills/llm-cli.md) — skill · 235 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
