Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 78% confidence
- Finding
L02 的技能名包含 “-zh”，整体文档也默认以中文描述输出流程，但前文未明确给用户提供语言选择或说明该技能仅限中文场景。根据规则，若技能在自然语言层面强制特定语言/locale 且无用户 opt-in，构成语言/locale 政策风险。
- Content
Security audit
Amazon 商品转视频脚本
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed Amazon product-analysis workflow that uses browser-based public page review/listing data to produce Chinese social video scripts, with no hidden code, persistence, credential use, or destructive behavior found.
Install if you want a Chinese Amazon product-to-video-script workflow. Be aware it may browse public Amazon pages using available browser automation and may create local markdown summaries/scripts; do not use it for private Amazon account data or login-required review scraping.
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
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 (4)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
The trigger condition is overly broad: any Amazon link plus general requests about short videos, scripts, or social media assets may activate this skill even when the user did not intend product-analysis behavior. That can cause misrouting, unnecessary browsing of third-party content, and execution of a multi-step workflow in unrelated conversations, increasing the chance of unintended data access or user-confusing actions.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The description states the skill outputs Chinese content by default, without indicating a fallback based on user language preference. This can override user expectations or platform language norms, causing misleading behavior, poor usability, and possible prompt-routing errors in multilingual environments.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The trigger set includes broad phrases such as “用户评论转视频脚本”, which can match general requests about turning user comments into video scripts without clearly requiring an Amazon product-analysis context. This can cause the skill to activate outside its intended scope, leading to incorrect routing, unintended data handling, or user confusion about what the skill is designed to do.
- Content