ClawHub
Back to skill

Security audit

Ora社媒主页搜索专家

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed social-media search integration that sends user search terms to fixed API endpoints and uses a local service token, with no evidence of hidden exfiltration or destructive behavior.

Install only if you are comfortable sending search keywords, company names, and domains to api.topeasychina.com. Treat OraAgent.key as a secret, avoid placing unrelated credentials there, and consider reducing token logging before use in sensitive environments.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill reads an AuthToken from a local file in the skills directory and automatically attaches it to outbound API requests. While this appears intended to support authenticated use of the social-media search service, it still grants the code access to a local secret that is not derived from per-request user input, increasing the blast radius if the skill or surrounding environment is compromised and creating a credential-handling risk beyond the user-facing search functionality.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill description does not warn users that their keywords, company names, and domains are transmitted to an external third-party service. This creates a privacy and data-handling risk, especially because business prospecting queries may contain sensitive commercial targets or research terms.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The README documents that user-supplied search terms, company names, and domains are sent via POST to external third-party endpoints and authenticated with a token, but it does not clearly warn users that their inputs will leave the local agent environment. In a lead-generation and social-account discovery skill, those queries may contain sensitive business intelligence, customer lists, or investigative targets, so the lack of disclosure can cause unintended data sharing and compliance/privacy issues.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.