Back to skill

Security audit

GenVR Skills

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a legitimate GenVR media CLI, but it needs Review because it loads broad local .env secrets, recommends mutable npx execution, and can overwrite arbitrary writable files through its output filename.

Install only if you are comfortable sending prompts and account identifiers to GenVR. Prefer a pinned reviewed version instead of unpinned npx, run it from a directory without unrelated .env secrets, and avoid using --filename with absolute paths, traversal paths, symlinks, or existing important files.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Findings (4)

T09 · Insecure Skill Coding Practices

Error
Location
bin/genvr.js:92
Finding

Unrestricted Output Path Allows Arbitrary File Overwrite

Content
View full analysis
{ https.get(url, (res) => { if (res.statusCode !== 200) { reject(new Error(`Failed to download: ${res.statusCode}`)); return; } const fileStream = fs.createWriteStream(dest); res.pipe(fileStream); fileStream.on('finish', () => { fileStream.close(); console.log(`Saved to ${path.resolve(dest)}`); resolve(); }); }).on('error', reject); }); } ``` ```js // Generate filename const cleanUrl = url.split('?')[0]; const ext = path.extname(cleanUrl) || '.png'; const timestamp = new Date().toISOString().replace(/[:.]/g, '-').slice(0, 19); const filename = args.filename || `${timestamp}-${args.category}${ext}`; await downloadFile(url, filename); ``` ### Technical Analysis The user-controlled `--filename` argument is passed directly to `fs.createWriteStream()` without validation or confinement to a designated output directory. Node.js permits absolute paths and relative traversal sequences such as `../`, so the destination can refer to any file writable by the current process. `fs.createWriteStream()` uses overwrite-compatible behavior by default. The implementation does not request exclusive creation, check whether the destination already exists, reject symbolic links, or verify that the resolved path remains inside an approved directory. Consequently, an existing file can be truncated and replaced. A symbolic link at the destination can also redirect the write to another file. The downloaded bytes originate fro ...[truncated 1314 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
bin/genvr.js:92
Finding

Remote-Controlled, Unbounded HTTPS Download

Content
View full analysis
{ https.get(url, (res) => { if (res.statusCode !== 200) { reject(new Error(`Failed to download: ${res.statusCode}`)); return; } const fileStream = fs.createWriteStream(dest); res.pipe(fileStream); fileStream.on('finish', () => { fileStream.close(); console.log(`Saved to ${path.resolve(dest)}`); resolve(); }); }).on('error', reject); }); } ``` ```js const resultData = resultResp.data || {}; const output = resultData.output; const url = Array.isArray(output) ? output[0] : output; if (!url) { console.log('Result data:', JSON.stringify(resultData, null, 2)); return; } const cleanUrl = url.split('?')[0]; const ext = path.extname(cleanUrl) || '.png'; const timestamp = new Date().toISOString().replace(/[:.]/g, '-').slice(0, 19); const filename = args.filename || `${timestamp}-${args.category}${ext}`; await downloadFile(url, filename); ``` ### Technical Analysis The output URL is supplied by the GenVR API and passed directly to `https.get()` without parsing or validating its hostname, port, or expected storage domain. If the API, an upstream service, or the API response is compromised, the CLI can be instructed to connect to an attacker-selected HTTPS endpoint, including HTTPS services reachable only from the user's network. The download also has no timeout, maximum response size, content-type validation, or file-type verification. The response is streamed until the server closes it. A malicious or faulty server c ...[truncated 1657 chars]
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Note
Location
bin/genvr.js:13
Finding

CLI Loads Every Variable from the Working Directory's .env File

Content
View full analysis
{ const [key, ...value] = line.split('='); if (key && value.length > 0) { const k = key.trim(); if (!process.env[k]) { process.env[k] = value.join('=').trim().replace(/^["']|["']$/g, ''); } } }); } } loadEnv(); ``` ### Technical Analysis The Skill requires only `GENVR_API_KEY` and `GENVR_UID`, but its loader reads the complete `.env` file from the current working directory and copies every parsed entry into `process.env`. This violates least privilege by exposing unrelated project secrets to the CLI process. Examples may include database passwords, cloud credentials, signing keys, and tokens for unrelated services. No audited code path was found that transmits these unrelated values, so confirmed secret exfiltration is not present. Nevertheless, broad loading unnecessarily expands the sensitive data available to the process and to any future or compromised code running in it. The selected file depends on `process.cwd()` rather than the package directory or an explicit configuration path. Running the command from a sensitive project therefore causes that project's complete `.env` file to be consumed automatically. ### Attack Path 1. A user or agent runs the CLI from a project directory containing a `.env` file. 2. The file contains the required GenVR values together with unrelated secrets. 3. `loadEnv()` reads the complete file before command dispatch. 4. Every parsed variable not already present is copied into `proc ...[truncated 757 chars]
Remediation
View remediation
0) { config[key] = rawValue.join('=').trim().replace(/^["']|["']$/g, ''); } } ``` ]]>

T08 · Insecure Dependencies

Note
Location
SKILL.md:23
Finding

Documentation Recommends Unpinned NPX Package Execution

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (20)

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · bin/genvr.js (reported line 13)May include surrounding context.

js
const https = require('https');
const { pipeline } = require('stream/promises');

// Helper: Minimal zero-dependency .env loader
function loadEnv() {
    const envPath = path.join(process.cwd(), '.env');
    if (fs.existsSync(envPath)) {

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · bin/genvr.js (reported line 15)May include surrounding context.

js
// Helper: Minimal zero-dependency .env loader
function loadEnv() {
    const envPath = path.join(process.cwd(), '.env');
    if (fs.existsSync(envPath)) {
        const content = fs.readFileSync(envPath, 'utf8');
        content.split('\n').forEach(line => {

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · .claude-plugin/plugin.json (reported line 8)May include surrounding context.

json
"author": "GenVR Research",
  "repository": "https://github.com/genvrresearch/genvr-skills",
  "entry": "SKILL.md",
  "icon": "https://api.genvrresearch.com/favicon.ico"
}

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The README instructs users to run the package via npx genvr-skills without pinning a specific version. That creates a supply-chain risk because the latest published package will be fetched and executed at runtime, so a compromised maintainer account or malicious update could result in arbitrary code execution in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

This example also uses unpinned npx genvr-skills, which causes users to download and execute whatever version is currently published. In agentic or automated environments, that increases exposure to package-replacement and malicious-update attacks.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · README.md (reported line 45)May include surrounding context.

Start Task:

bash
curl -X POST https://api.genvrresearch.com/v2/generate \
  -H "Authorization: Bearer $GENVR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill documents capabilities that require access to environment variables and outbound network communication, but it does not declare an explicit tool scope such as permissions or allowed-tools. This creates a governance gap where consumers may not understand the full execution and data-access surface, increasing the risk of unintended secret exposure or unreviewed external requests.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · README.md (reported line 45)May include surrounding context.

md
Generate images, videos, and perform various media processing tasks using the GenVR API. This standalone Node.js toolkit provides a pure CLI for interacting with GenVR's asynchronous task pipeline.

**IMPORTANT:** Most operations require a `GENVR_API_KEY` and a `GENVR_UID`. Get them from [api.genvrresearch.com](https://api.genvrresearch.com/obtain-api).

## Usage

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 17)May include surrounding context.

md
Generate images, videos, and perform various media processing tasks using the GenVR API. This standalone Node.js toolkit provides a pure CLI for interacting with GenVR's asynchronous task pipeline.

**IMPORTANT:** Most operations require a `GENVR_API_KEY` and a `GENVR_UID`. Get them from [api.genvrresearch.com](https://api.genvrresearch.com/obtain-api).

## Usage

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

The skill instructs users to run npx genvr-skills without pinning a specific package version. This allows whatever package version is current at execution time to be fetched and run, creating a supply-chain risk if a future release is compromised, malicious, or materially changes behavior.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

This invocation uses npx genvr-skills without an exact version, so execution depends on the latest published package at runtime. That makes the skill vulnerable to upstream package compromise or unexpected code changes being executed in the user's environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

Running an unpinned npx package executes remote code that may differ over time from what was reviewed. In a skill that also uses API credentials, this raises the stakes because a compromised package could exfiltrate secrets or misuse the account.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

The status command still relies on unpinned npx execution, exposing users to supply-chain compromise even for seemingly low-risk operations. Any execution path can run arbitrary package lifecycle or CLI code before performing the requested action.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

This example again executes the package via unpinned npx, inheriting the same remote code execution and supply-chain risks. Because the tool is designed for networked media generation, a compromised package could silently send prompts, credentials, or local metadata to attacker-controlled endpoints.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

The image-to-video example invokes an unversioned package from the registry at runtime. This is dangerous because users may trust documentation examples and execute whatever code is currently published, including a compromised release with access to environment secrets and network connectivity.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The generate command sends user-supplied prompts, task metadata, and a UID to a remote API, but the CLI provides no explicit warning or consent notice at the point of transmission. In a media-generation tool, prompts may contain sensitive business, personal, or proprietary data, so silent transmission increases the risk of unintended data disclosure.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
85% confidence
Finding

The comment instructs users to execute the package via 'npx genvr' without a pinned version, which can fetch and run whatever package version is current at execution time. This creates a supply-chain risk: a compromised publisher account, malicious new release, or dependency hijack could cause unreviewed code to run on the user's system.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The markdown instructs users to export GENVR_API_KEY and GENVR_UID and later use them with remote API calls, but it does not include any user-facing warning about protecting these credentials or that prompts/data will be sent to an external GenVR service. For markdown files, missing warnings about privacy- or system-relevant behavior should be flagged when the description omits them.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The comments state this package is 'primarily a collection of Python-based CLI tools' and instruct using 'npx genvr' to run them, while the file itself exports a Node.js package descriptor and the manifest context describes a standalone Node.js CLI. This is an active documentation contradiction about what the skill is, rather than merely omitted detail.

Content

No source excerpt is available for this finding.