AI电商专家｜电商文生视频

• Outcome: Review
• Latest audit: Aug 23, 2026 · 9:48 AM UTC
• Version: 1.0.0

Overview

The skill mostly matches its ecommerce video purpose, but it runs an unpinned external npm tool with the user's token and full environment, so it should be reviewed carefully before installation.

Install only if you trust the IMIVA/@infimind npm package and the remote service. Prefer pinning the npm package version, run it from a shell without unrelated secrets, use a limited/revocable MCP token, and confirm credits and local media uploads before creating tasks.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (7)

subprocess module call

Medium

• Category: Dangerous Code Execution

Content

env = os.environ.copy()
    env["MCP_TOKEN"] = token
    env["API_URL"] = os.environ.get("IMIVA_API_URL", os.environ.get("API_URL", DEFAULT_API_URL))
    proc = subprocess.Popen(
        [npx, "-y", PACKAGE],
        stdin=subprocess.PIPE,
        stdout=subprocess.PIPE,

• Confidence: 91% confidence
• Finding: proc = subprocess.Popen( [npx, "-y", PACKAGE], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, bufsize=1,

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill documents use of shell commands, environment variables, and local file paths, but it does not declare corresponding permissions or clearly constrain when those capabilities may be used. In an agent setting, hidden or undeclared capability use can lead to unexpected local file access, token exposure, or command execution beyond what the user understood when invoking the skill.

Vague Triggers

Medium

• Confidence: 79% confidence
• Finding: The description advertises the skill for broad ecommerce, listing, and social-commerce workflows rather than tightly limiting it to text-to-video generation. Overly broad invocation scope increases the chance that an orchestrator routes unrelated requests into a skill that has shell, file, and token-handling instructions, creating unnecessary exposure and capability creep.

Vague Triggers

Medium

• Confidence: 76% confidence
• Finding: The applicable-user section is broad enough to match many normal Chinese ecommerce requests from multiple roles, which can cause ambiguous or excessive invocation. In context, this matters because the skill includes operational guidance involving external tools, task creation, and potentially local/secret-bearing resources, so unnecessary activation expands attack surface and risk of unintended actions.

Missing User Warnings

Medium

• Confidence: 86% confidence
• Finding: The tool silently executes an external npm package in a subprocess, which can perform network access and arbitrary actions without any explicit user confirmation at invocation time. In a skill context, this reduces user awareness and increases the chance that unreviewed third-party code runs unexpectedly.

Missing User Warnings

Medium

• Confidence: 80% confidence
• Finding: The script requires an `MCP_TOKEN` and uses it to invoke a network-capable MCP service, but the code itself provides no runtime disclosure about what remote service will be contacted or that authenticated actions will occur. This can surprise users and leads to accidental token use in a context that executes external tooling.

Env Variable Harvesting

High

• Category: Data Exfiltration

Content

def invoke(method: str, params: dict | None = None) -> dict:
    npx, token = require_runtime()
    env = os.environ.copy()
    env["MCP_TOKEN"] = token
    env["API_URL"] = os.environ.get("IMIVA_API_URL", os.environ.get("API_URL", DEFAULT_API_URL))
    proc = subprocess.Popen(

• Confidence: 93% confidence
• Finding: os.environ.copy()

VirusTotal

VirusTotal findings are pending for this skill version.

Static analysis

No suspicious patterns detected.