ClawHub
Back to skill

Security audit

项目管理API

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a project-management GraphQL helper, but it grants broad local tools and external API mutation/callback authority without tight scoping or data-handling limits.

Review this skill before installing. Use a narrowly scoped API key, avoid broad task-planning prompts unless you intend API actions, and do not provide a callback URL unless you trust the destination and understand what project data may be sent.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Vague Triggers

Medium
Confidence
79% confidence
Finding
The activation language is broad enough to match many generic project-management requests, which can cause the skill to trigger in situations beyond the user's intended scope. In an agent environment, overbroad activation increases the chance that write/exec-capable tooling is selected for routine tasks, expanding attack surface and enabling unintended external API actions.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill declares read, write, and exec capabilities and describes command execution behavior without prominent user-facing warnings about system or file impact. In an agent context, that creates a meaningful risk of unintended file modification or shell execution because users may invoke what appears to be a simple project-management integration without understanding the local side effects.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
Documenting callback_url support without a privacy or data-handling warning creates a clear exfiltration risk because agent outputs or project data may be sent to arbitrary external endpoints. In this skill's context, the data could include issue content, comments, team metadata, or operational results, making silent transmission particularly sensitive.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.