Seedance Video Creation

• Outcome: Review
• Latest audit: May 28, 2026 · 11:35 AM UTC
• Version: 1.0.0

Overview

The core video generator is understandable, but it adds under-declared Feishu file sharing and auto-opening behavior that users should review before installing.

Install only if you are comfortable sending prompts and selected images to Volcengine under your ARK_API_KEY. Treat the Feishu instructions as a separate external sharing action: confirm the exact file, recipient/chat, and sensitivity before any send, and protect Feishu app credentials. The publisher should disclose these data flows more clearly, remove or split the Feishu guide, and make file opening an explicit opt-in.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

os.system() or os exec-family call

High

• Category: Dangerous Code Execution

Content

# Open on macOS
                    if sys.platform == "darwin":
                        os.system(f'open "{filepath}"')
                except Exception as e:
                    print(f"Download failed: {e}", file=sys.stderr)

• Confidence: 95% confidence
• Finding: The code invokes a shell via os.system() to open the downloaded file on macOS. Even though the path is quoted, shell execution is unnecessary here and creates avoidable command-invocation risk if path handling changes or quoting is bypassed; it also performs a local side effect beyond simple task management.

Context-Inappropriate Capability

Medium

• Confidence: 88% confidence
• Finding: The Feishu/OpenClaw file-sending feature expands the skill from video generation into outbound file delivery, which is a separate data-flow capability. This increases exfiltration risk because generated or downloaded videos could be forwarded to another external service outside the user’s original expectation.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The file documents Feishu file upload and messaging flows, including reading a local file and transmitting it to external Feishu/CDN endpoints, which is outside the declared purpose of a video-generation skill. Expanding scope to cross-tool messaging and exfiltration-adjacent behavior increases risk because users may invoke the skill expecting only video generation, while the documentation normalizes onward transfer of generated or local files to third-party services.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The skill automatically opens a downloaded video on macOS, which is outside the narrowly described purpose of generating, querying, and downloading videos. Triggering local application behavior without explicit consent increases the attack surface and can surprise users in security-sensitive environments.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The documentation does not prominently warn that prompts, local images, and possibly derived media are transmitted to ByteDance/Volcengine. Because local files may be base64-encoded and uploaded, users could unintentionally send sensitive images or confidential prompt contents to a third party.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: This section explicitly shows opening a local file and uploading its contents to Feishu/CDN, but it does not clearly warn that local data will leave the workspace and be sent to an external service. That omission can cause unintended data disclosure, especially in an agent environment where users may not realize a local path could contain sensitive or non-generated content.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The documentation references Feishu app_id, app_secret, and app_access_token without labeling them as sensitive secrets that must not be exposed, logged, or embedded in prompts or examples. In operational practice, such omission can lead to credential mishandling and unauthorized use of the Feishu application if those values are leaked.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: Local image files are read, base64-encoded, and then included in requests to a remote service, but the tool does not provide an explicit warning at the point of use that local content will be uploaded off-host. In a CLI handling user media, this can lead to unintended disclosure of sensitive images.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The tool sends prompts and media references to a remote API endpoint but does not prominently warn users that their content is transmitted to an external service. Because the skill's purpose is remote video generation, network transmission is expected, but lack of explicit disclosure still creates privacy and compliance risk.

Static analysis

No suspicious patterns detected.