ClawHub
Back to skill

Security audit

Android 输入法窗口稳定性

Security checks for vulnerabilities and agentic risk

Overview

The scanners report no security findings from AIG and one low-probability output-handling issue from Skillspector; the verdict is benign with medium confidence, though some referenced files were unavailable for review.

The verdict is benign, but that does not guarantee the skill is risk-free. Skillspector scanned both listed components, yet marked its analysis partial because several referenced files were missing; AIG completed with no findings. Consider whether you are comfortable installing with those review limits.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (1)

The skill requires PR descriptions in Chinese without offering a language choice.

Medium
Category
Output Handling
Confidence
98% confidence
Finding

The instruction “中文 PR 正文” explicitly mandates Chinese-language PR descriptions. The file provides neither user opt-in nor a documented justification for this language restriction; writing the skill itself in Chinese does not establish consent.

Content

No source excerpt is available for this finding.