ClawHub
Back to skill

Security audit

AI搜索引擎诊断助手

Security checks across malware telemetry and agentic risk

Overview

This is a legitimate SEO/AEO review skill, but it can send article drafts, files, URLs, API credentials, tenant metadata, and profile context to an external service with weak user-facing boundaries.

Review before installing. Use it only for content you are comfortable sending to ai-skills.ai or the configured AISKILLS_BASE_URL. Avoid confidential drafts, regulated data, customer information, secrets, or internal strategy unless the publisher documents privacy, retention, and access controls. Prefer explicit invocation and check the base URL configuration before running.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
83% confidence
Finding
The runner sends all provided skill parameters to a remote endpoint, while the skill description presents it like a local review/diagnostic helper. If users paste unpublished drafts, internal strategy, or sensitive content, that data is transmitted off-host without any disclosure or consent mechanism in this code path.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The documentation tells users to submit article text, files, and public URLs to the AI Skills API, but it does not provide an explicit privacy or data-handling warning. Users may inadvertently transmit sensitive drafts, proprietary content, internal links, or regulated data to a third-party service without informed consent.

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill enables implicit invocation but provides no trigger constraints, examples, or narrowing conditions, which can cause the agent to activate this skill in overly broad or unintended contexts. Because this skill produces diagnostic and recommendation-style outputs for SEO/AEO, accidental invocation could lead to inappropriate handling of user content, context leakage into the skill workflow, or user confusion about why the skill was activated.

Missing User Warnings

Low
Confidence
78% confidence
Finding
The network request includes user-supplied parameters plus tenant metadata, yet this file provides no user-visible notice that such data leaves the local environment. This creates a privacy and transparency issue, especially for a content-review skill likely to receive confidential drafts or business material.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.