ClawHub
Back to skill

Security audit

asyncguard

Security checks for vulnerabilities and agentic risk

Overview

AsyncGuard is mostly a local code scanner, but its license validation can execute commands from a crafted license token, so it needs Review before installation.

Treat this as Review. The local scanner and optional hooks are understandable, but do not install or configure paid-tier license tokens unless you trust the publisher and token source. Review lefthook changes before enabling hooks, and prefer a fixed version that parses JWT JSON through stdin or files instead of embedding decoded token text in interpreter command strings.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (2)

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
This script can create, append to, and remove `lefthook.yml` in the current repository, which modifies project configuration from an analyzer entry point. In an agent/skill context, repository mutation is more dangerous because users may expect read-only analysis behavior, and these writes can persist code-execution hooks into future developer workflows.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The hook installation path writes or appends to repository configuration without a clear up-front warning that files will be modified. In a security-sensitive skill environment, silent persistence mechanisms are risky because they can change future commit behavior and establish unexpected execution paths.

Static analysis

No suspicious patterns detected.