Security audit

Realchinese

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local Chinese text detector and rewriter with no evidence of hidden access, persistence, or data exfiltration.

Install only if you want a Chinese-language local writing helper; expect heuristic AI-trace scoring and style rewriting, but do not rely on the advertised 29-detection claim or pip-install claim without additional packaging support.

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

MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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 (6)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 99% confidence
Finding: The description emphasizes a '真人化引擎' that rewrites Chinese AI text into more human-like style and supports 8 platform-specific rewriting modes. However, the supplied code only analyzes input text and computes an AI-likelihood score with a report. It does not modify text, rewrite style, or provide any platform-specific transformation features. Additionally, while the description mentions 29 AI-trace detections, this code chunk defines only 15 actual detection metrics (p1–p15), despite comments claiming 29. This is a material description-versus-behavior mismatch, with the code's primary purpose being detection rather than humanization.
Content: No source excerpt is available for this finding.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 97% confidence
Finding: The description partially matches the rewriting portion: the script clearly supports 8 Chinese platform-oriented rewrites and attempts to make text feel more human via template removal, slang/emoji insertion, hooks, CTA, tags, and some structural variation. However, a major part of the declared purpose is unsupported: there is no code for '29 AI-trace detection' or any detector, classifier, scoring, reporting, or analysis features. The packaging claim ('pip install后直接使用') is also not evidenced by this code chunk, which is just a module/script. Because the declaration prominently advertises a detection engine in addition to rewriting, while the actual code only implements rewriting utilities, this is a material description-behavior mismatch.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The manifest and heading describe the skill exclusively as a Chinese text humanization engine and present usage entirely in Chinese, with no indication that users can choose another language or locale. This creates a natural-language locale constraint without documented opt-in or justification as a region-specific compliance tool.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This code file contains user-facing natural-language strings that present the tool entirely in Chinese, including the product description and command help. Under the policy, forcing a specific language without user opt-in or a clearly documented justified locale constraint is a natural-language policy violation.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The module description and usage examples state that the detector is specifically for Chinese text and all user-facing output strings are in Chinese. This imposes a language/locale constraint without any opt-in or alternative, which matches the policy category for forced language behavior.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The module description and example usage present the skill as exclusively producing Chinese-platform style text, and the public API later restricts platform choices to Chinese platforms only. This is a natural-language locale constraint with no opt-in or alternative language path documented, which fits the policy category for forced language/locale behavior.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.