ClawHub
Back to skill

Security audit

Contract Check

Security checks for vulnerabilities and agentic risk

Overview

This contract-review skill appears purpose-aligned and local-only, with a manageable privacy consideration around saved review settings.

Before installing, treat the saved configuration as sensitive business data. Use it only in a trusted workspace, avoid storing confidential negotiation thresholds unless needed, and remove the local config when it should no longer be reused.

SkillSpector

Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • 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)

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill persists company-specific legal and commercial risk preferences to a local file without explicitly warning the user that the data will be stored and retained. This can expose sensitive internal business rules, negotiation thresholds, and contract positions to unintended local access or future reuse beyond the user's expectations.

Session Persistence

Medium
Category
Rogue Agent
Content
metadata:
  permissions:
    - file:read
    - file:write
  behavior:
    network: none
    telemetry: none
Confidence
93% confidence
Finding
The skill has file write permission and instructs persistent storage of contract-review configuration in the user's workspace, creating session persistence across interactions. Persistent state is not inherently malicious here, but it increases risk because sensitive company policy data can remain on disk, be reused in later sessions without renewed consent, or be accessed by other local processes or skills sharing the environment.

Static analysis

No suspicious patterns detected.