Security audit

血友病小红书内容创作专家

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Xiaohongshu content-generation helper that fetches user-provided content and creates images/text without hidden code, credential use, or destructive behavior.

Install this if you want a Chinese Xiaohongshu post generator. Review any medical-health claims before publishing, provide only links you are comfortable having processed, and be aware that generated images are intended to be saved under the workspace output path.

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

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill accepts very broad inputs such as arbitrary topics, article links, or video content without clear scoping, validation boundaries, or restrictions on supported sources. That increases the chance of unsafe downstream processing, unexpected data extraction, or misuse of the skill for unreviewed external content, especially since it automatically pulls content from websites before generating output.

Natural-Language Policy Violations

Medium

Confidence: 98% confidence
Finding: The skill hardcodes Chinese-language output in its image prompt template rather than adapting to the user's requested language. While this is not a classic security flaw, it can override user intent and create integrity or usability issues in multi-language contexts, especially if users expect faithful transformation of source material.

Natural-Language Policy Violations

Low

Confidence: 72% confidence
Finding: The instruction to '保持"血友乐天派"账号风格' hard-codes a specific account voice/style, which can function as a locale/content policy constraint for all outputs without asking whether the user wants that persona. This may inappropriately override user intent unless clearly limited to a dedicated, account-specific skill context.

Static analysis

No suspicious patterns detected.