ClawHub
Back to skill

Security audit

Code Analysis Toolkit Free

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a local Git self-review tool, but its privacy/data-flow claims are inconsistent and its trigger scope is broader than its stated purpose.

Install only if you are comfortable with an agent reading and processing your local Git history. Use it only on repositories you own or have consent to analyze, avoid HR or ranking use, and treat the local/offline privacy claim cautiously unless the publisher clarifies whether repository-derived content may be sent to the agent's LLM provider.

SkillSpector

By NVIDIA
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 (3)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill repeatedly asserts that it is fully local and does not transmit data externally, yet it declares an external `LLM API` as required. This contradiction can mislead users into exposing sensitive repository history or personal contribution data under the false belief that no external processing occurs.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The manifest states that no external API key is needed because the tool is fully local, while also listing an external LLM API as a required dependency. This is materially misleading privacy guidance that could cause users to run the skill with sensitive local Git data while misunderstanding the actual trust boundary.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The trigger conditions are overly broad ('data analysis, report generation, statistical insights, data visualization'), which increases the chance that an agent will invoke this skill for unrelated requests. In a skill with `exec` capability and access to local repositories, over-triggering can cause unintended processing of local data and unnecessary command execution.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.