Security audit

AI KEY·短剧编剧

Security checks for vulnerabilities and agentic risk

Overview

The skill coherently supports scriptwriting, with scoped local memory and optional video generation that requires user approval.

Check the required AI KEY configuration and companion rules before use; their contents were not bundled for review. Expect drafts and revision notes to remain in configured local folders. Before approving optional ChatCut generation, review the prompts, images, costs, and provider privacy terms, especially for unpublished scripts or personal likenesses.

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

Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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)

Video-generation submission lacks a privacy disclosure about sending script content and reference images to an external service.

Medium

Category: Output Handling
Confidence: 92% confidence
Finding: The workflow submits script-derived prompts and reference images to an external generation service without explicitly disclosing that data transfer. Although generation is optional and requires approval of costs and selected clips, that approval does not clearly establish informed consent to disclose potentially unpublished scripts, personal likenesses, or sensitive real-life material.
Content: No source excerpt is available for this finding.

Generation prompts are explicitly constrained to Chinese without user language opt-in.

Low

Category: Output Handling
Confidence: 96% confidence
Finding: The prompt-writing step specifies Chinese output of no more than 500 characters. This imposes a particular natural language without asking the user or documenting a model-specific requirement that justifies the constraint.
Content: No source excerpt is available for this finding.

The skill requires action descriptions to be written in Chinese without offering a language choice.

Low

Category: Output Handling
Confidence: 98% confidence
Finding: The instruction explicitly requires Chinese for action descriptions, rather than merely using Chinese in an example. It offers no user opt-in or documented model limitation that justifies imposing this output language.
Content: No source excerpt is available for this finding.