ClawHub
Back to skill

Security audit

seedance-creator

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Seedance/dreamina media-generation helper that may use a logged-in CLI account and upload user-selected media, but I found no hidden or destructive behavior.

Install this only if you intend to use the dreamina CLI with a logged-in account. Review generated prompts and file paths before execution, avoid sensitive local media, and prefer explicit /seedance use so it does not run for unrelated image or video tasks.

A.I.G

Vulnerability Patterns
  • 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
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Session Persistence

Medium
Category
Rogue Agent
Confidence
78% confidence
Finding

The skill is designed to operate with browser-based OAuth credentials persisted by the dreamina CLI in the user's home directory, and it encourages repeated CLI-driven media operations once installed and logged in. In combination with broad activation, this creates a session-persistence risk where future invocations may silently act with an already-authenticated account, enabling unintended third-party uploads or generation tasks without a fresh trust check.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: seedance-creator
description: This skill should be used when the user asks to "generate video prompts", "create Seedance prompts", "write video descriptions", mentions "Seedance", "seedance", "即梦", "即梦平台", "视频提示词", "视频生成", "AI视频", "短剧", "广告视频", "视频延长", "生成图片", "文生图", "图生图", "图生视频", "文生视频", or discusses video prompt engineering, AI video generation, or Seedance 2.0 workflows. It also handles requests to create, edit, or manipulate images and videos using the dreamina CLI tool.
version: 5.0.1

# 安全须知

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The activation description is extremely broad and includes generic image/video creation phrases, making the skill likely to trigger outside a narrow Seedance-specific context. Because the skill can invoke a local CLI, read user-supplied file paths, and interact with persisted OAuth credentials, over-triggering increases the chance of unintended tool use and accidental data exposure.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

Mandating Chinese-only prompts without user opt-in is not a classic security flaw, but it is a policy constraint that can silently transform user input and reduce transparency about what is actually sent to the external service. In a tool-execution skill, hidden rewriting can cause users to approve one request while another semantically altered prompt is transmitted, increasing misuse and consent risk.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.