ClawHub
Back to skill

Security audit

代码质量检查基础版

Security checks across malware telemetry and agentic risk

Overview

The skill mostly matches a local code-quality checker, but its broad activation wording and conflicting local-only versus external API language need review before installation.

Install only if you want an agent to read project files and run local shell commands for code-quality checks. Confirm before running broad scans, avoid printing full secret values in reports, and do not create the optional Git pre-commit hook unless you explicitly want that persistent local behavior. Treat the API/network language as unclear until the publisher clarifies whether any external service is contacted.

SkillSpector

Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The skill is presented as a local-only code quality checker, but later claims API integration, network-dependent behavior, and broader runtime capabilities. This inconsistency can mislead users and orchestrators about the skill’s trust boundary, causing it to be invoked in contexts where network access or external data flows are not expected.

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The documentation explicitly states that no external API key is needed and that checks run locally, while nearby sections describe external API integration and network-dependent operation. Security-relevant contradictions like this undermine informed consent and can conceal unintended data exposure paths, especially when source code contents may be processed during analysis.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger wording is broad enough to overlap with ordinary development, debugging, code generation, and deployment requests. Overbroad triggers can cause the skill to auto-activate in unrelated contexts, increasing the chance that `exec`-capable behavior is introduced when the user did not specifically request this skill or its command execution patterns.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The quick-start invocation example is so generic that it could match many normal project-help conversations rather than a deliberate request to run this specific skill. In an agent ecosystem, generic trigger examples raise the likelihood of unintended skill selection and unnecessary command execution against the user’s workspace.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.