Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 84% confidence
- Finding
描述和全文内容均以中文呈现，且没有说明可根据用户偏好切换语言或提供多语言选项。按规则，若技能默认强制特定语言而无用户选择，属于语言/区域策略方面的自然语言政策风险。
- Content
Security audit
Company Values
Security checks for vulnerabilities and agentic risk
Overview
This skill is a Chinese-language business advice template for drafting company values and does not request tools, data access, persistence, or execution authority.
Installers should know this skill may activate for general Chinese-language discussions about company culture, values, hiring, or remote teams. It appears safe, but narrowing the activation language would make it less likely to steer unrelated conversations.
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
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- 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)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
The trigger keywords include very broad terms such as “价值观” and “公司文化”, which can cause the skill to activate in many ordinary business conversations that are not specifically asking for this workflow. Over-broad activation is risky because it can override better-matched skills, inject unsolicited guidance, or route user intent incorrectly, though this particular skill’s content is non-privileged and low-risk.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 89% confidence
- Finding
The read_when conditions are broadly phrased and lack exclusion criteria, so the orchestrator may load this skill whenever a user mentions hiring, values, or remote culture in a general way. This increases the chance of unintended invocation and response steering, but the danger is limited by the skill’s advisory-only business content.
- Content
Static analysis
No suspicious patterns detected.