Seedance 商品视频工作室｜AI-HIVE

• Outcome: Pass
• Latest audit: Aug 25, 2026 · 11:26 AM UTC
• Version: 1.0.0

Overview

This skill is a disclosed AI-HIVE product-video workflow that uses local scripts, user-selected media, and an API key in ways that match its stated purpose.

Install only if you intend to use AI-HIVE for product-video work. Review generated commands before running them, provide only media you are authorized to upload, and be aware that init stores an AI-HIVE API key locally while generation commands may consume paid credits after confirmation.

SkillSpector

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill advertises executable workflows and shell/network/file/env capabilities but does not declare permissions, which weakens least-privilege controls and reduces user/operator visibility into what the skill can access. In this context, the included commands can read environment variables, access local files, invoke external APIs, and perform local processing, so an undeclared-capability skill could be invoked with broader authority than expected.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 90% confidence
• Finding: The documented purpose is narrowly framed as a product-video studio workflow, but the behavior described by analysis includes broader generic chat, generic image generation, account/model querying, standalone media upload, and general ffmpeg editing utilities. This mismatch is dangerous because users and policy layers may authorize the skill for a limited e-commerce workflow while it actually exposes more general-purpose capabilities that can touch account data, upload arbitrary media, or process local files beyond the stated scope.

Vague Triggers

Medium

• Confidence: 80% confidence
• Finding: The trigger phrases are broad enough to match many ordinary marketing, advertising, and e-commerce requests, increasing the chance this skill is selected outside its intended narrow use case. Overbroad activation becomes more risky here because the skill can lead to network calls, media uploads, local file operations, and potentially billable generation actions, so accidental invocation can expand data exposure and operational impact.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The default prompt hard-wires invocation of the skill via "$seedance-product-video-studio-ai-hive" in a generic way that can cause the agent to activate this skill even when the user's request is only loosely related. Because this skill is capable of producing executable workflows and runnable AI-HIVE commands, unintended activation can expand scope, trigger external API-oriented actions, or steer outputs in ways the user did not explicitly request.

Natural-Language Policy Violations

Medium

• Confidence: 84% confidence
• Finding: The default prompt forces a "production-ready Chinese workflow" regardless of the user's language preference or consent. This can misalign outputs with user intent, hide important warnings or compliance details from non-Chinese-speaking users, and make review of generated commands or business claims less reliable.

Static analysis

No suspicious patterns detected.