Security audit

CreBee | 多平台视频解析下载

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed CreBee helper for downloading user-provided video links to the local Downloads folder, with no evidence of hidden persistence, exfiltration, or destructive behavior.

Install this only if you use CreBee and are comfortable with the agent downloading videos through the local CreBee client, consuming CreBee API quota, and saving completed files to your system Downloads folder after you ask it to download a link.

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

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Missing User Warnings

High

Confidence: 97% confidence
Finding: The skill explicitly tells the agent to suppress important operational details from the user and only provide a minimal prompt before initiating a download that writes a file to the local system. Hiding preconditions, destination, limitations, and runtime characteristics reduces informed consent and can cause users to trigger local file creation without understanding what will happen.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The skill explicitly instructs the agent to perform a download and save a file to the user's local Downloads directory as part of the normal flow, and even says not to repeat confirmation once the user asks to download. That creates a real consent/transparency issue because it normalizes a local filesystem side effect without requiring the agent to clearly warn where the file will be written or to verify the save action at execution time.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill states that a user's request to download is sufficient authorization and instructs the agent to execute immediately without confirmation, even though the action causes a file to be written locally. This weakens safeguards against mistaken, ambiguous, or maliciously induced downloads and increases the chance of unwanted files being stored on the system.

Natural-Language Policy Violations

Medium

Confidence: 97% confidence
Finding: This troubleshooting guide is entirely written in Chinese and directs agent behavior and user-facing phrasing in Chinese, but it does not indicate that Chinese is optional, user-selected, or required for a region-specific compliance reason. Under the language/locale policy, forcing a specific language without opt-in is a natural-language policy violation.

Static analysis

No suspicious patterns detected.