Back to skill

Security audit

Next Video Gen

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its media-generation purpose, but bundled scripts create review-level risk through a legacy URL injection flaw and an unrestricted authenticated HTTP helper.

Review before installing. Prefer the Node implementation and avoid running scripts/seedance-gen.sh with untrusted video or audio URLs until the interpolation bug is fixed. The maintainer should restrict scripts/http-request.js to HTTPS Volcengine Ark hosts before attaching DOUBAO_API_KEY. Treat prompts and media URLs as data sent to Volcengine, use a scoped API key, and remember generated files are saved locally by default.

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/seedance-gen.sh:209
Finding

Arbitrary JavaScript Execution Through Unsafely Interpolated Media URLs

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/http-request.js:7
Finding

API Key Disclosure Through an Arbitrary-Destination HTTP Helper

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (25)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared description presents the skill as a media generation capability. However, the supplied code chunk does not perform any video/image generation or call Volcengine/Seedance APIs. Its primary purpose is installation and environment setup: filesystem writes, dependency detection, OpenClaw skills directory discovery, and API key configuration persistence. These are materially different from the declared runtime functionality, so this code chunk is not accurately represented by the description.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README describes prompts, images, and videos being processed via the Volcengine Ark API but does not clearly warn users that their inputs will be transmitted to a third-party service. In a skill designed for agent-driven media generation, users may provide sensitive prompts or private media, so missing disclosure creates a meaningful privacy and data-handling risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README instructs users to send prompts, images, and videos to a third-party cloud generation service using an API key, but it does not clearly disclose that user-supplied content will be transmitted off-device to Volcengine/Ark. In an agent skill context, users may provide sensitive prompts or media assuming local processing, so the lack of an explicit data-sharing warning creates a meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README instructs users to configure an API key and use cloud-based media generation, but it does not clearly disclose that prompts, images, and videos may be transmitted to a third-party service and that generated outputs are stored locally on disk. In a skill that handles user-supplied media and potentially sensitive prompts, this omission can lead to unintended data exposure, privacy issues, and unsafe handling of confidential content.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill declares capabilities that require environment access and shell execution, but it does not explicitly scope or constrain those permissions. That increases the attack surface because consumers and hosting agents cannot clearly enforce least privilege, and future script behavior may gain broad execution access without transparent review.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill clearly sends user prompts and user-provided image/video/audio URLs to Volcengine Ark APIs, but it does not prominently disclose that user content is transmitted to a third-party service for processing. This can create privacy, confidentiality, and compliance risk if users submit sensitive prompts or private media under the assumption processing is local.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · bin/cli.js (reported line 48)May include surrounding context.

js
console.log('  node bin/cli.js --version  ' + dim('# show version'));
  console.log('');
  console.log(bold('Options:'));
  console.log('  -y, --yes        ' + dim('Non-interactive mode. Auto-detect skills dir, skip prompts.'));
  console.log('  --path <dir>     ' + dim('Install to a specific directory (used with -y).'));
  console.log('');
  console.log(bold('What this installer does:'));

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · bin/cli.js (reported line 159)May include surrounding context.

js
}
  }

  // Silent mode: create default path without asking
  if (opts.silent) {
    const defaultDir = path.join(home, '.openclaw', 'skills');
    console.log(yellow('  ⚠  No skills directory found, creating: ') + defaultDir);

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · bin/cli.js (reported line 159)May include surrounding context.

js
}
  }

  // Silent mode: create default path without asking
  if (opts.silent) {
    const defaultDir = path.join(home, '.openclaw', 'skills');
    console.log(yellow('  ⚠  No skills directory found, creating: ') + defaultDir);

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · bin/cli.js (reported line 261)May include surrounding context.

js
required: true,
      installHint: {
        darwin: 'brew install curl  (or: xcode-select --install)',
        linux: 'sudo apt install curl  (or: sudo yum install curl)',
        win32: 'Download from https://curl.se/windows/',
      },
    },

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · bin/cli.js (reported line 270)May include surrounding context.

js
required: true,
      installHint: {
        darwin: 'brew install curl  (or: xcode-select --install)',
        linux: 'sudo apt install curl  (or: sudo yum install curl)',
        win32: 'Download from https://curl.se/windows/',
      },
    },

Session Persistence

Medium
Category
Rogue Agent
Confidence
90% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · bin/cli.js (reported line 367)May include surrounding context.

js
console.log('');
  console.log(bold('Next steps:'));
  console.log('  1. ' + dim('Ensure DOUBAO_API_KEY is set in your environment'));
  console.log('     ' + dim('export DOUBAO_API_KEY=your_key  (or add to .zshrc/.bashrc)'));
  console.log('  2. ' + dim('Load the skill in your agent and start generating videos'));
  console.log('     ' + dim('"Generate a 5-second 720p video of a sunset over the ocean"'));
  console.log('');

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The documentation tells users they can provide external image/video/audio URLs as reference media, but it does not warn that those URLs and referenced content will be transmitted to a third-party service provider for processing. In a media-generation skill, users may mistakenly submit private or sensitive assets, creating privacy, confidentiality, and compliance risks rather than direct code-execution risk.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This utility accepts an arbitrary URL from the command line and automatically attaches the DOUBAO_API_KEY as a Bearer token to every request. In the context of a video/image generation skill, that creates a confused-deputy capability: a caller can direct authenticated requests to unintended endpoints, potentially exfiltrating the API key to attacker-controlled services or using the tool for actions outside the skill’s stated purpose.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script can send arbitrary payloads to user-supplied destinations while silently including an API credential from the environment, with no confirmation or visibility guardrails. In an agent skill, this increases the risk of unintended data transfer and secret exposure because the helper normalizes outbound authenticated requests without constraining what data is sent or where it goes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The script reads the sensitive environment variable DOUBAO_API_KEY to authenticate API requests, but the file provides no warning about credential usage beyond requiring it to be set. Under the code-file criteria, sensitive credential access should have some user disclosure such as a comment, docstring, or explicit warning about how the key will be used.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script submits generation payloads to external endpoints using http-request.js, including PROMPT text and potentially image, video, or audio URLs. Although network transmission is core functionality here, the file itself does not clearly disclose that user-supplied content is sent to a third-party service, which is a privacy-relevant behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README states that generated media is saved under ~/Videos/next-video-gen/ but does not clearly warn that outputs will be written to persistent local storage. This can expose sensitive generated content to other local users, backups, sync tools, or forensic recovery if users assume generation is ephemeral.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

The document is entirely written in Traditional Chinese and labels itself as the Traditional Chinese variant. While alternate language files are linked, this file itself does not state any user-selectable locale behavior or justify a locale restriction, which can be read as a fixed-language experience in the skill documentation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The file presents all user-facing instructions and API guidance exclusively in Chinese, with no indication that the skill is intentionally China/Chinese-locale-specific or that alternative language support is available. Under the stated policy, forcing a single language without user opt-in can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The usage/comment says the tool handles -d '{"json"}', implying the next argument is direct JSON input. In reality, the code first attempts to read process.argv[5] as a filesystem path and only falls back to inline JSON on failure, so the documented behavior contradicts the primary operation actually performed.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

This code emits a Chinese-only mode label (文生图) in its user-facing output. The file does not offer a language selection mechanism or explain that the skill is intentionally region-specific, so it creates a natural-language locale policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The modeInfo labels and submission status output use hard-coded Chinese strings for multiple modes. Because the skill provides no language preference setting or justification for a Chinese-only experience, this is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The progress log 视频生成中... (已等待 ... 秒) is user-facing and fixed to Chinese. Without user opt-in or documentation of a required locale, this violates the language/locale policy requirement.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The script creates an output directory and automatically downloads generated media into local files, but there is no upfront disclosure that running the skill will write files under $HOME/Videos/next-video-gen by default. For code files, file writes should have some user-facing warning, log, or explanatory comment.

Content

No source excerpt is available for this finding.