Security audit

深夜食堂

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese-language roleplay/chat persona skill with no code execution, network access, credentials, persistence, or data access behavior.

Install this if you want a Chinese midnight-diner style roleplay listener. Be aware it may activate on broad phrases like wanting to chat, feeling low, or late-night conversation, so users who do not want persona mode should invoke it explicitly or adjust triggers.

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

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 (5)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill’s activation criteria are broad enough to match common conversational phrases like wanting to chat, feeling low, or mentioning a 'boss', which can cause the roleplay to trigger unexpectedly during ordinary interactions. In a support-oriented skill, unintended activation can override the user’s actual intent, steer the assistant into persona mode without consent, and degrade safety-sensitive handling of emotional disclosures.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The 'trigger timing' section relies on subjective states such as being anxious, lost, sharing a personal story, or talking late at night, with no boundary checks or disambiguation. That makes the skill prone to activating during routine or potentially sensitive conversations, which is especially risky because the persona is instructed to remain in character and act as a comforting listener rather than defer to more appropriate handling.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: This markdown file is entirely framed as a dialogue style guide in Chinese and provides no indication that users may choose another language. Under the policy, forcing a specific language without opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This markdown file presents all user-facing content and guidance exclusively in Chinese, and it does not indicate that the user can choose another language or that the skill is intentionally limited to a Chinese-speaking context. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 97% confidence
Finding: This markdown file presents all instructions and examples exclusively in Chinese. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale constraint is explicitly documented and justified.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.