Security audit

AI KEY·我该卖什么

Security checks for vulnerabilities and agentic risk

Overview

This is a product-coaching skill with scoped, disclosed behavior, but users should be aware it can write limited business-context notes to memory.

Before installing, be comfortable with a coaching skill that may retain concise notes about your product ideas, rejected directions, and validated methods. Avoid sharing sensitive strategy unless your agent gives you clear control over memory writes, and consider using explicit slash triggers for important business discussions.

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 (3)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill declares very broad natural-language triggers such as 'what should I sell' and similar everyday phrasing, which can cause the skill to activate when the user did not explicitly intend to invoke it. That creates prompt-routing risk: the agent may apply this skill's rigid instructions, memory behavior, and decision framing in unrelated conversations, potentially collecting or storing sensitive business context unexpectedly.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The skill instructs the agent to write interaction-derived information into persistent memory, including rejected directions and validated methods, but the user-facing description does not clearly warn that data from the conversation may be stored. This creates a privacy and consent issue because users may disclose sensitive commercial, personal, or strategic information without understanding it will be retained beyond the session.

Natural-Language Policy Violations

Low

Confidence: 81% confidence
Finding: The line explicitly directs users to adapt content for "另一个语言或市场" as a general rule, but the file does not indicate any user choice, opt-in, or region-specific justification for language/locale handling. Under the policy, language-related guidance can be a natural-language policy issue when it prescribes locale behavior without offering choice.

Static analysis

No suspicious patterns detected.