Media Gen (wangdong233) MCP Server

media-gen-mcp-server npm v0.19.0

Published by wangdong233 — no publish provenance, so origin is unverified, but the source is public: the repository link below is self-declared yet readable, so you can inspect the code before adopting it.

All-in-one image-generation MCP server for Claude Code: AI imagery/video (Agnes AI + Zhipu) + structured local rendering (diagrams, charts, formulas, cards, icons, QR codes). Provider- and engine-pluggable.

Trust grade
A
92/100
Last scanned get badge →
Trust
A · 92/100
Adoption risk for you: the threat score, then adjusted down for blast radius, publisher verification and how much the scan could see. Deterministic; every point is auditable.
Capability
High
Blast radius if it went rogue — what the server’s tools could reach. Independent of trust.
Coverage
Source
How much the scan could actually inspect. Shallow coverage is stated, never hidden.
Share this Trust Score
𝕏 Share LinkedIn Reddit
A Why this grade threat 99 − adoption risk = 92/100

The grade answers one question — how safe is this server for you to adopt — so it is computed in two auditable stages. Nothing below is an opinion or an LLM's guess; every line is a real term the deterministic engine applied, and the same input always yields the same number.

1. Threat score — 100 − 1.2 = 99. What the published surface and source actually contain:

PointsWhat was foundCategory
−1.2 Hardcoded Google API key in server code MTC-SRC-008 exfiltration

2. Client adoption risk — 99 − 7 = 92. Three small, subtract-only factors that reflect your risk in adopting it — a clean scan proves less on a powerful, unverified or barely-inspectable package, so the grade says so plainly:

PointsAdoption-risk factor
−6 capability blast radius (high) — client exposure if the model is manipulated
−1 publisher verification (public source) — no provenance, but the source is public and inspectable

Capability observations and info notes are shown under Findings but never scored. Open any row's finding below for the file, line and evidence behind a deduction.

Findings 7

high Dynamic evaluation of a non-literal value (dist/providers/flow.js)MTC-SRC-010

In the server's implementation (`dist/providers/flow.js:391`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: .value; } async eval(expr, timeoutMs) { try { return await this.evalOnce(expr, timeoutMs);

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/providers/flow.js

high Dynamic code execution in server code (dist/providers/flow.js)MTC-SRC-001

In the server's implementation (`dist/providers/flow.js:391`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: .value; } async eval(expr, timeoutMs) { try { return await this.evalOnce(expr, timeoutMs);

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/providers/flow.js

high Shell/command execution in server code (dist/browser-pool.js)MTC-SRC-002

In the server's implementation (`dist/browser-pool.js:2`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: { execFile } from "node:child_process"; import { promisify } from "node:util"; import path from "node:path"; /** * brow

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/browser-pool.js

high Shell/command execution in server code (dist/checks/output-checker.js)MTC-SRC-002

In the server's implementation (`dist/checks/output-checker.js:22`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: spawnSync } from "node:child_process"; import path from "node:path"; import { createRequire } from "node:module"; // ──

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/checks/output-checker.js

high Shell/command execution in server code (dist/render-video.js)MTC-SRC-002

In the server's implementation (`dist/render-video.js:1`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: spawnSync } from "node:child_process"; import fs from "node:fs"; import path from "node:path"; import os from "node:os"

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/render-video.js

medium Hardcoded egress to an external endpoint (dist/providers/flow.js)MTC-SRC-003

In the server's implementation (`dist/providers/flow.js:297`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: const res = await fetch(`http://${CDP_HOST}:${this.port}/json/list`, { signal: AbortSignal.timeout

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/providers/flow.js

low Hardcoded Google API key in server code (dist/providers/flow.js)MTC-SRC-008

A hardcoded Google API key (a client API key (restricted by referrer/quota, not a private secret)) appears in `dist/providers/flow.js:37`. Verify whether this is a real credential; if so, remove and rotate it.

Evidence: Google API key: AIza…(redacted)

Fix: Remove the secret, rotate it, and load credentials from the environment or a secret store.

Location: server dist/providers/flow.js

Tools 23

Each tool and what it can reach — statically extracted from the published source.

  • flow_statusingests untrusted input
  • generate_iconingests untrusted input
  • generate_imageingests untrusted input
  • get_videoingests untrusted input
  • analyze_chartno sensitive capability
  • create_videono sensitive capability
  • describe_imageno sensitive capability
  • extract_image_metano sensitive capability
  • extract_pdfno sensitive capability
  • extract_tableno sensitive capability
Show 13 more tools ↓
  • extract_textno sensitive capability
  • generate_cardno sensitive capability
  • generate_chartno sensitive capability
  • generate_diagramno sensitive capability
  • generate_formulano sensitive capability
  • generate_interactive_diagramno sensitive capability
  • generate_nested_diagramno sensitive capability
  • generate_qrcodeno sensitive capability
  • get_pdfno sensitive capability
  • list_modelsno sensitive capability
  • list_vision_capabilitiesno sensitive capability
  • render_svgno sensitive capability
  • render_videono sensitive capability

What this scan could not see

Versions 7

Scan history per published version. The engine is deterministic — the same version always yields the same score, so a changed score means the package itself changed.

VersionScoreFindingsEngineScanned
v0.19.0 latest A 92/100 7 1.13.0 2026-09-07
v0.17.1 A 92/100 6 1.13.0 2026-09-01
v0.15.0 A 92/100 6 1.13.0 2026-08-31
v0.14.0 A 93/100 2 1.13.0 2026-08-27
v0.13.1 A 93/100 2 1.12.1 2026-08-12
Show 2 more versions ↓
v0.12.1 A 93/100 2 1.12.1 2026-07-28
v0.11.0 A 100/100 1 1.5.0 2026-07-22

Embed this score

Show this server's live Trust Score in your README, docs or website. The badge is served straight from the registry and updates automatically after every rescan — no API key needed. It links back to this page, so anyone who sees the grade can also read the findings behind it instead of taking a number on faith.

MCP Trust Score: A · 92/100
Markdown (GitHub README)
[![MCP Trust Score](https://mcptrustchecker.com/registry/media-gen-mcp-server/badge.svg)](https://mcptrustchecker.com/registry/media-gen-mcp-server)
HTML
<a href="https://mcptrustchecker.com/registry/media-gen-mcp-server"><img src="https://mcptrustchecker.com/registry/media-gen-mcp-server/badge.svg" alt="MCP Trust Score" height="20"></a>
Prefer shields.io styling? Point it at https://mcptrustchecker.com/registry/media-gen-mcp-server/badge.json via https://img.shields.io/endpoint?url=…

Verify this score yourself

The score above is reproducible: the same package version always yields the same result. Run it locally or over the free API — no account, no LLM, fully deterministic.

npx mcptrustchecker scan media-gen-mcp-server --online

Use the free API → How scoring works

Other implementations of Media Gen 2

Independent packages implementing the same tool, scanned with the same engine. Compare all 3 side by side →

More in Browser Automation