Back to skill

Security audit

Seedance 图生视频工作室|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly aligned with AI-HIVE video generation, but it contains credential and file-write risks that users should review before installing.

Install only if you are comfortable using an AI-HIVE API key and uploading selected media to that service. Avoid setting AI_HIVE_BASE_URL or --base-url unless you fully trust the endpoint, use a limited/rotatable API key, and review generated commands before running them. Treat downloaded results cautiously because the current script should sanitize filenames before writing files.

Vulnerability Patterns
  • 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
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:61
Finding

API Credential Disclosure Through an Unrestricted Custom Base URL

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:379
Finding

Arbitrary File Write Through Unsanitized Server-Controlled Download Filenames

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (8)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared description promises a fairly full Seedance图生视频 workflow, including accepting image-based inputs, producing specific video-planning outputs, and optionally performing AI-HIVE OpenAPI operations to generate materials. The supplied code does something much narrower: it parses a few text arguments and writes a static JSON brief containing metadata, aliases, required facts, artifact labels, and draft status. While blueprint creation loosely relates to pre-production planning, the primary declared capability is a comprehensive image-to-video studio process, and the actual code does not implement the central behaviors or external integrations described. Therefore the description materially overstates what this code chunk actually does.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill advertises and demonstrates code paths that use environment variables, local files, network access, and shell execution, but it does not declare any explicit tool scope or permissions boundary. In an agent setting, this increases the chance the skill is granted broader capabilities than intended, enabling unintended file access, outbound requests, or command execution if the skill is invoked automatically.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The default prompt uses broad wording ('turn my request') combined with implicit invocation, which can cause the skill to activate on loosely related user requests rather than clear opt-in for Seedance image-to-video tasks. Because this skill can produce runnable AI-HIVE commands and workflow steps, unintended activation could steer users into external API usage, cost-incurring actions, or media-generation flows they did not explicitly request.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Forcing output to Chinese in the default prompt without user opt-in can override user expectations and reduce transparency about what actions or commands are being proposed. In a workflow that may generate runnable commands and operational checklists, language coercion can increase the risk of user misunderstanding, mistaken execution, or inability to review safety-relevant details.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is presented as a narrow Seedance image-to-video studio, but the file exposes a much broader generic AI client with text chat, image generation, model enumeration, account inspection, and arbitrary media upload. This capability mismatch violates least privilege and increases the attack surface: any caller who can invoke the skill may gain access to functions unrelated to the advertised purpose, enabling policy bypass or unintended data/API usage.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The module docstring and user-facing guidance are written entirely in Chinese, and later CLI help and prompts continue this pattern, which imposes a specific language/locale on users. The file does not indicate that language is configurable or optional, so it violates the language-choice policy for natural-language content.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

Exposing user-info and wallet/billing visibility is not necessary for a Seedance image-to-video workflow and reveals account metadata beyond the skill's declared scope. Even if read-only, this can leak sensitive operational information such as balances or account identifiers and may aid reconnaissance or misuse of the linked AI Hive account.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The skill name, argument description, defaults, and search aliases are presented in Chinese-only text such as "图生视频工作室" and "待确认", with no indication that other languages are supported. This can violate a language/locale policy when users are not given a choice or informed that the skill is Chinese-language only.

Content

No source excerpt is available for this finding.