ClawHub
Back to skill

Security audit

Deepseek聊天

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a DeepSeek chat helper, but it asks for broad read, write, and command execution powers without clearly limiting when those powers may be used.

Review this carefully before installing. It may be appropriate only if you intentionally want a DeepSeek-related automation skill with local file and command capabilities; otherwise prefer a narrower chat/API skill that does not request write and exec authority.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The skill is branded and described as a DeepSeek chat/API tool, but later documentation expands its behavior to file processing, external API integration, and system command execution. That mismatch is dangerous because users or orchestrators may invoke it with lower trust assumptions, while it actually has broader system-impacting capabilities that could enable unexpected file modification or command execution.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The 'Use when' guidance is extremely broad, covering generic efficiency, automation, batch processing, and workflow optimization. In an agent ecosystem, this can cause over-selection of a skill that has powerful permissions, increasing the chance it is invoked for tasks beyond a user's reasonable expectation or without adequate scrutiny.

Vague Triggers

Medium
Confidence
88% confidence
Finding
Repeated ambiguous activation language broadens the apparent scope of the skill and reinforces selection for many unrelated productivity tasks. Because the skill advertises access to read, write, and exec tools, ambiguous routing materially raises the risk of unintended high-privilege use.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The documentation mentions file writing, external API calls, and command execution, but does not provide a prominent upfront warning that the skill can affect the local system and external services. This is dangerous because users may treat it as a simple chat integration while it can perform side-effecting actions that change files, run commands, or send data externally.

Static analysis

No suspicious patterns detected.