Security audit

cui-skill

Security checks for vulnerabilities and agentic risk

Overview

This is a text-only Chinese nagging/coaching persona skill with no code execution, data access, persistence, or hidden system changes, though its broad triggers and harsh tone may be annoying if enabled unintentionally.

Install this only if you want a Chinese, deliberately pushy productivity coach. Expect occasional activation on casual procrastination phrases and a harsh tone; avoid enabling it in contexts where that style would be stressful or inappropriate.

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

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill description and trigger model are broad enough that ordinary mentions of procrastination or casual requests could activate a strong coercive persona unexpectedly. Because the skill is designed to escalate pressure and impersonate an aggressive '催' assistant, accidental activation can degrade user experience, override conversational intent, and pressure users in contexts where the behavior is inappropriate.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The skill strongly fixes the persona and output style in Chinese without any user language preference check, which can cause unwanted language switching and reduce usability or comprehension for non-Chinese users. While not a direct security exploit, it is a behavioral safety issue because it can override user expectations and interfere with reliable communication.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The smart trigger logic maps very common phrases like '等会儿', '马上', and '明天再说' to activation states, which makes accidental triggering likely in normal conversation. In this skill, accidental triggering is more concerning because the generated behavior includes escalating pressure, ridicule, and repeated follow-up, creating a meaningful risk of unwanted harassment-like responses.
Content: No source excerpt is available for this finding.