Atlas / MCP servers / antonylu0826 / Zenku

ZenkuBLOCK

Name: Zenku
Author: antonylu0826

mcp/antonylu0826/zenku

MCP server by antonylu0826 · ai agents · application builder · database management · Audited 2026-10-08

AI-first No-Code engine for building production-ready data apps via conversation. Features Multi-Agent architecture, dynamic UI rendering, and built-in Undo/Time-Machine. Supports SQLite, Postgres, and MSSQL.

Verdict

BLOCK

Grade

Trust score

69 /100

Exposed tools

4 2r · 2w · 0d

Transport

streamable-http

License

NOASSERTION

Stars

Overview

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

Build Production-Ready Data Apps via Conversation.

繁體中文 | Documentation

Zenku is an AI-first, open-source No-Code Engine designed for building enterprise-grade data applications. Instead of generating static code, Zenku uses a sophisticated Multi-Agent Architecture to dynamically evolve your database schema, UI views, and business logic in real-time through natural language.

Chat-to-App Workflow: Transform "I need a CRM" into a live system (Schema -> CRUD UI -> Dashboard) in seconds.
Specialist Agent System: Dedicated agents for Schema, UI, Logic, Query, and Testing, coordinated by a central Orchestrator.
Enterprise DB Support: Native support for SQLite, PostgreSQL, and Microsoft SQL Server (MSSQL).
Versatile View Engine: Dynamic rendering for Kanban, Gantt, Calendar, Timeline, Dashboard, and more.
Real-time Logic: A powerful Business Rules Engine for automation, validation, and third-party webhooks (e.g., n8n).
Reliability & Undo: A built-in Design Journal acting as a "Time Machine," allowing you to undo any AI change instantly.
MCP Ready: Full Model Context Protocol support, enabling external AI tools to control your Zenku instance.

🧠 Architecture: The Multi-Agent Orchestrator

Unlike simple "Code Gen" wrappers, Zenku separates reasoning from execution through specialized roles:

graph TD
User[User]  Orch[Orchestrator Agent]
Orch  Schema[Schema Agent]
Orch  UI[UI Agent]
Orch  Query[Query Agent]
Orch  Logic[Logic Agent]
Orch  Test[Test Agent]

🏗️ Architecture

Zenku uses a Monorepo (npm workspaces) structure:

1. `@zenku/server` (The Brain)

Read from source at commit 5fdcbd16a862OBSERVED · 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. Replace the environment placeholders with a token scoped to the least it needs.

claude-code

claude mcp add web --env ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} --env GEMINI_API_KEY=${GEMINI_API_KEY} --env OPENAI_API_KEY=${OPENAI_API_KEY} --env OPENROUTER_API_KEY=${OPENROUTER_API_KEY} -- npx -y @zenku/[email protected]

claude-desktop

{
  "mcpServers": {
    "web": {
      "command": "npx",
      "args": [
        "-y",
        "@zenku/[email protected]"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "${ANTHROPIC_API_KEY}",
        "GEMINI_API_KEY": "${GEMINI_API_KEY}",
        "OPENAI_API_KEY": "${OPENAI_API_KEY}",
        "OPENROUTER_API_KEY": "${OPENROUTER_API_KEY}"
      }
    }
  }
}

Exposed tools (4)

2 read · 2 write · 0 destructive.

Tool	Risk	Description
`assess_impact`	read	Assess impact of destructive schema changes. Must call this tool before executing drop_column, rename_column, change_type, or drop_table. Reports affected interfaces, rules, record count, and foreign key dependencies.
`query_data`	write	Query data and answer statistics questions. Can only execute SELECT queries.
`set_translations`	write	Register or update translation entries for user-defined content ($key → display text per locale). Call this after creating schema or views when the user
`undo_action`	read	Undo previous operations. Call when user says

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 (12)

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

packages/server/src/db/mssql-adapter.ts:332

EXEC('ALTER TABLE [' + @tbl + '] DROP CONSTRAINT [' + @con + ']');

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94

packages/web/src/components/blocks/TableView.tsx:421

const blob = new Blob(['' + csv], { type: 'text/csv;charset=utf-8;' });