Security audit

China Collective Memory

Security checks for vulnerabilities and agentic risk

Overview

This is a read-only Chinese cultural memory reference skill with some dataset quality issues but no evidence of hidden execution, credential handling, persistence, or unsafe authority.

Install only if you want a Chinese-language reminiscence and cultural-history reference corpus. Treat entries as prompts to support conversation, not as authoritative history, personal testimony, or technical repair instructions; watch for occasional placeholders or fictional/mapped records in generated output.

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (17)

Natural-Language Policy Violations

Medium

Confidence: 95% confidence
Finding: This markdown skill file is entirely written in Chinese and provides all prompts, labels, and invocation content in that language without indicating any language choice or opt-in. Under the stated policy, forcing a specific language or locale without offering user choice is a natural-language policy violation.

Natural-Language Policy Violations

Medium

Confidence: 94% confidence
Finding: This markdown file presents all headings, prompts, and descriptions exclusively in Chinese, but it does not state that the skill is intended only for Chinese-speaking users or provide any opt-in language selection. Under the policy rule for language/locale, forcing a specific language without user choice can be a natural-language policy violation.

Natural-Language Policy Violations

Medium

Confidence: 94% confidence
Finding: SQP-3 applies to all file types and covers natural-language language or locale policy violations. This markdown file presents all instructions, prompts, and content in Chinese only, with no indication that the skill is region-specific or that users can opt into another language, which can conflict with organizational language-choice expectations.

Natural-Language Policy Violations

Medium

Confidence: 86% confidence
Finding: This markdown file is entirely written in simplified Chinese and repeatedly frames prompts for Chinese-speaking users without any indication that language selection is optional. Under the policy rule, a skill should not implicitly force one language or locale unless it offers user choice or clearly documents a justified regional scope.

Intent-Code Divergence

Medium

Confidence: 91% confidence
Finding: The manifest positions the skill as a search/browse corpus of memory anchors for reminiscence work, implying plausible prompts grounded in lived or commonly mediated memory. Entries such as "东方红一号卫星发射成功" describe direct sensory details like launch roar and launch-site soil that conflict with the accompanying cultural background saying ordinary people learned of the event via broadcast or newspapers, making the prompt inconsistent with its intended reminiscence use.

Description-Behavior Mismatch

Medium

Confidence: 95% confidence
Finding: The manifest describes a corpus for Chinese collective-memory anchors oriented to reminiscence and cultural history. This file includes entries like "《烟雨濛濛》专辑" and later items such as "邓丽君歌曲", "魔方发明", and "香港廉政公署成立", which are not clearly justified as Chinese collective-memory anchors from the file’s own framing and broaden the corpus beyond the declared scope.

Natural-Language Policy Violations

Medium

Confidence: 95% confidence
Finding: This markdown file presents its title and all subsequent user-facing content in Chinese, and there is no indication that the language is optional or that users may choose another locale. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.

Natural-Language Policy Violations

Medium

Confidence: 92% confidence
Finding: This markdown file is written entirely in Chinese and all probe questions and descriptions assume Chinese-language interaction, but there is no indication that the skill is intentionally limited to Chinese-speaking users or that users can opt into another language. Under the policy rule, forcing a specific language without opt-in is a locale/language policy concern.

Natural-Language Policy Violations

Medium

Confidence: 90% confidence
Finding: SQP-3 applies to all file types and covers language or locale policy violations. This markdown file is entirely written in Chinese and repeatedly addresses a Chinese-speaking audience, but nowhere indicates that the content is intentionally China-specific or that other language options are available.

Intent-Code Divergence

Medium

Confidence: 98% confidence
Finding: This file is presented as a decade-organized reference of collective-memory anchors, but the entry at L13458-L13468 is marked as a '虚构/映射' concept and its cultural background explicitly says 'User Interest Insert', indicating it was injected to reflect a current user's interests rather than document an authentic 2006-era memory anchor. That documentation contradicts the apparent intent of the file as a historical reference corpus for browsing real anchors across 1935-2025.

Ae4

Medium

Category: analysis-evasion
Confidence: 80% confidence
Finding: Suspicious Unicode normalization or mixed-script content

Natural-Language Policy Violations

Medium

Confidence: 96% confidence
Finding: SQP-3 applies to all file types and covers natural-language language/locale policy violations. This markdown file consistently presents all headings, descriptions, and prompts in Chinese, and nowhere offers an alternative language option or states that the content is intentionally limited to a Chinese-language audience.

Natural-Language Policy Violations

Medium

Confidence: 86% confidence
Finding: This markdown file is written entirely in Chinese from the title onward, with no indication that other languages are supported or that the Chinese-only scope is intentional and justified. Under the policy rule for language or locale constraints, this can amount to forcing a specific language without user opt-in.

Natural-Language Policy Violations

Low

Confidence: 79% confidence
Finding: The working note instructs agents to 'use it directly when working with Chinese speakers,' which imposes a language behavior based on user group rather than explicit user preference. Under the policy rule, language constraints should either offer choice or be clearly documented as an opt-in or justified regional limitation.

Intent-Code Divergence

Low

Confidence: 95% confidence
Finding: The file contains an explicit placeholder entry marked as not for production use, yet it remains embedded among normal records. In a retrieval or generation pipeline, this can leak internal notes into user-facing output, confuse downstream parsers, and undermine trust in dataset integrity.

Context-Inappropriate Capability

Low

Confidence: 84% confidence
Finding: L08999-L09008 provides step-like guidance for replacing plugs and splicing electrical wires, which is a practical repair capability rather than a memory/search anchor for reminiscence or cultural-history storytelling. While the file is mostly memory prompts, this entry crosses into actionable household electrical work that is not necessary for the declared purpose.

Context-Inappropriate Capability

Low

Confidence: 81% confidence
Finding: At L13460-L13469, the content centers on replacing resistors/capacitors and soldering repairs to a radio, emphasizing DIY technical repair skill. That level of repair-oriented capability is not directly justified by a skill whose purpose is to search and browse memory anchors for reminiscence and intergenerational storytelling.

Static analysis

No suspicious patterns detected.