ClawHub
Back to skill

Security audit

Text Punctuation Restoration/文本标点恢复与断句段落划分

Security checks for vulnerabilities and agentic risk

Overview

This is a simple Chinese text-formatting skill with some disclosed content changes users should notice, but no evidence of malicious or high-impact behavior.

Install only if you are comfortable with the skill converting Traditional Chinese to Simplified Chinese and possibly creating a separate titled version of the text. Review outputs before using them for legal, archival, quoted, or style-sensitive material.

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 (3)

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill states a highest-priority rule that the original text must not be changed and only punctuation may be added, then immediately introduces an exception requiring Traditional-to-Simplified conversion. This contradiction weakens the reliability of the skill contract and can lead agents to modify user content beyond the declared transformation boundary, especially for quoted, legal, literary, or archival text where script form is semantically important.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

Forcing Simplified Chinese conversion without user consent changes user-provided content in a way that may alter meaning, provenance, stylistic intent, or compliance requirements. In this skill's context, users expect formatting restoration, not orthographic normalization, so the undisclosed mandatory conversion exceeds the least-privilege expectation for a text-processing skill.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill manifest and top-level description present the capability as punctuation restoration and sentence breaking, but the body of the skill adds a second-stage behavior that generates hierarchical summary titles and writes output to a new file. This scope expansion can cause downstream agents or users to trust the skill with a narrower data-transformation role than it actually performs, increasing the risk of unintended content generation, data reshaping, or unauthorized file-output behavior.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.