Security audit

Seedance 视频提示词工业化生成

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Seedance video-prompt workflow that creates a Markdown deliverable, with the main caution being ordinary file-write control.

Install only if you want a Chinese Seedance-focused workflow that may save large generated prompt files. Before running it, specify the output path and ask the agent to confirm before overwriting an existing file; otherwise the behavior is consistent with its stated purpose.

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

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (6)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The trigger conditions are broad and primarily keyword-based, so the skill may activate for loosely related requests such as general video/story help rather than an explicit request to run this workflow. In an agent setting, unintended invocation can cause the model to apply heavyweight transformations, request or infer extra structure, and proceed into file-generation behavior without clear user intent.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill explicitly directs writing output to a .md file and only reporting the file path, total groups, and duration, with no user confirmation or safety check. In an agent environment with filesystem/tool access, this can create or overwrite files unexpectedly, hide substantive output from the user, and make side effects occur without an explicit approval step.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The entire skill is written as a Chinese-only workflow and requires final output in a Chinese-named Markdown artifact, but it does not state that language choice is optional or user-selected. This creates a natural-language locale policy concern because the skill implicitly enforces a specific language without opt-in.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill explicitly instructs the agent to write generated content directly into a Markdown file in the working directory, despite no metadata or manifest indicating filesystem access is required or constrained. This is dangerous because it normalizes side effects without user confirmation, enabling unauthorized file creation/overwrite and making it easier for a prompt-injection-style skill to alter local workspace state.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill directs the agent to write output straight to a Markdown file without warning, confirmation, or opt-in from the user. In an agent setting this is risky because it can trigger unexpected persistent changes, overwrite existing files under predictable naming rules, and bypass the principle that side-effecting actions should be explicit and consented to.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 74% confidence
Finding: The positioning at L007 describes the skill as a speed-first, lightweight prompt-generation workflow. However, L813 requires the final file to include not only image/video prompts but also a 台词核对表 and 字段自检表, adding audit/reporting deliverables that contradict the claimed lightweight scope rather than merely leaving details unspecified.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.