Atlas / MCP servers / peakmojo / Agentic Client

Agentic ClientBLOCK

Name: Agentic Client
Author: peakmojo

mcp/peakmojo/agentic-client

MCP server by peakmojo · anthropic api · anthropic claude · aws bedrock · Audited 2026-10-08

A standalone agent runner that executes tasks using MCP (Model Context Protocol) tools via Anthropic Claude, AWS BedRock and OpenAI APIs. It enables AI agents to run autonomously in cloud environments and interact with various systems securely.

Verdict

BLOCK

Grade

Trust score

69 /100

Exposed tools

3 3r · 0w · 0d

Transport

—

License

Apache-2.0

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Current Features

Included a basic agent dashboard
Run standalone agents with tasks defined in JSON configuration files
Support for both Anthropic Claude and OpenAI models
Session logging for tracking agent progress

Run Dashboard Web

cd dashboard
npm i
npm run dev

Dashboard URL: http://localhost:3000 API Documentation: http://localhost:3000/api-docs

https://github.com/user-attachments/assets/c98be6d2-0096-40f2-bd78-d3fb256fec83

Installation

Clone the repository

Set up dependencies:

uv sync

Create an agent_worker_task.json file

Here is an example configuration file:

{
"task": "Find all image files in the current directory and tell me their sizes",
"model": "claude-3-7-sonnet-20250219",
"system_prompt": "You are a helpful assistant that completes tasks using available tools.",
"verbose": true,
"max_iterations": 10
}

Run the agent:

uv run agentic_mcp_client/agent_worker/run.py

Configuration

The project requires a config.json file in the root directory to define the inference server settings and available MCP tools. Here's an example configuration:

{
"inference_server": {
"base_url": "https://api.anthropic.com/v1/",
"api_key": "YOUR_API_KEY_HERE",
"use_bedrock": true,
"aws_region": "us-east-1",
"aws_access_key_id": "YOUR_AWS_ACCESS_KEY",
"aws_secret_access_key": "YOUR_AWS_SECRET_KEY"
},
"mcp_servers": {
"mcp-remote-macos-use": {
"command": "docker",
"args": [
"run",

Read from source at commit 894939e282f2OBSERVED · 2026-10-08

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code

claude mcp add dashboard -- npx -y [email protected]

claude-desktop

{
  "mcpServers": {
    "dashboard": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ]
    }
  }
}

Exposed tools (3)

3 read · 0 write · 0 destructive.

Tool	Risk	Description
`Configurations`	read	API endpoints for configuration files
`Jobs`	read	API endpoints for managing agent worker jobs
`Sessions`	read	API endpoints for accessing agent worker sessions

Trust audit

BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	PASS
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (5 observation(s))
Network: declared (5 observation(s))
Shell: declared (1 observation(s))
Dependencies: not all pinned
Secrets in source: none-found

Findings (7)

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

dashboard/app/api/jobs/route.ts:189

const childProcess = exec(

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table