ClawHub
Back to skill

Security audit

视频生成 Video Generate

Security checks for vulnerabilities and agentic risk

Overview

The skill’s video-generation workflow clearly discloses its cloud uploads, account-key storage, and software installation, with no material security concern found.

Use media you are comfortable uploading to dLazy and account for generation-credit costs. The CLI can save your organization API key locally; the skill also documents environment-variable authentication and key revocation. The supplied static scan reported no findings, and the advisory description and whitespace concerns were downgraded after inspecting both language versions.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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
Findings (3)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

The first pipeline matches the declared image-to-video purpose, and its initial image generation is a supporting step. However, the other examples introduce TTS and standalone image-processing workflows, particularly a batch super-resolution pipeline that produces no video. These extend beyond the declared purpose. The chunk also uses fixed model commands rather than demonstrating automatic model selection, though that omission alone would not establish a mismatch in a partial code sample.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The first pipeline fits the declared video-generation purpose, with image generation serving as a supporting step. The other pipelines produce speech or enhanced images without generating video, exposing capabilities beyond the declared scope. The supplied chunk also does not demonstrate automatic model selection, although that absence alone is not evidence of a mismatch. No unrelated resource access or triggers are shown.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL-cn.md (reported line 152)May include surrounding context.

md
## 错误处理

| Code | 错误类型                         | 示例信息                                                                                                                 |
| ---- | -------------------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| 401  | 未授权 (API Key缺失或无效)       | `ok: false, code: "unauthorized", message: "API key is missing or invalid"`                                                              |
| 501  | 缺少必填参数                     | `error: required option '--prompt <prompt>' not specified`                                                               |

Static analysis

No suspicious patterns detected.