ClawHub
Back to skill

Security audit

Code Analysis Toolkit

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed Git analysis workflow, but it asks to analyze sensitive repository and author data while making conflicting local-only and LLM/API data-flow claims.

Install only if you are comfortable with an agent analyzing the selected repositories and author metadata. Before use, confirm whether the agent or LLM provider may receive repository history, verify the actual `src.main` code that will run, restrict scans to intended repositories, and protect or delete `.code-analysis` logs/reports when no longer needed.

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 (1)

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The skill claims it is fully local and does not transmit data externally, yet elsewhere requires an LLM API and references LLM service timeouts. This contradiction can mislead users into providing sensitive repository history or personal data under false privacy assumptions, creating a real confidentiality and compliance risk even if no overt exfiltration logic is shown in the document.

Static analysis

No suspicious patterns detected.