Security audit

Doubao AI Image Generator

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says: it uses a logged-in Doubao browser session to generate and return images, with no hidden code or persistence found.

Install only if you are comfortable sending image prompts and generated content to Doubao through a logged-in account. Avoid sensitive prompts, and periodically clean temporary generated image files from the workspace if privacy matters.

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)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger text is broad enough to route essentially any image-generation request to a skill that automates a logged-in third-party website. That can cause user prompts to be sent off-platform without clear consent and may invoke account-scoped browser actions in situations where a narrower, safer tool would be more appropriate.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill instructs the agent to use a logged-in browser session against doubao.com but does not warn that prompts and related content will be transmitted to a third party under the user's account context. In this context, omission of that disclosure is risky because users may unknowingly expose sensitive text or operate through a personal account session.

Natural-Language Policy Violations

Medium

Confidence: 94% confidence
Finding: The prompt guidance and examples push Chinese-language optimization terms without offering locale or language choice. While not a classic security flaw, it can misrepresent user intent and cause unintended transmission or transformation of prompts to a different language, which is more problematic in a third-party browser-automation skill handling user content.

Missing User Warnings

Low

Confidence: 87% confidence
Finding: The requirements mention a workspace directory for temporary file storage, and later steps download generated images locally, but the skill does not warn that files will be written to disk. This creates a modest confidentiality and hygiene risk because generated images may persist in temporary storage longer than the user expects.

Static analysis

No suspicious patterns detected.