ClawHub
Back to skill

Security audit

company-profiling-zhcn

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed PatSnap-based pharmaceutical company profiling workflow with a limited web-search fallback, and I did not find hidden installation, persistence, destructive behavior, or unrelated data access.

Before installing, understand that this skill depends on a PatSnap API key and will send relevant company, drug, target, patent, and clinical queries to PatSnap MCP. It may also use general web search when PatSnap data is insufficient or when you ask for recent updates, so avoid submitting confidential company or deal information unless you are comfortable with those external queries.

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 (1)

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The skill explicitly authorizes general web search after MCP queries are exhausted, expanding data access beyond the declared PatSnap-only company profiling scope. This is dangerous because it can cause uncontrolled retrieval from untrusted external sources, increasing risks of prompt injection, unreliable data incorporation, privacy leakage via search queries, and inconsistent security boundaries for a skill that users may expect to be limited to curated MCP data.

Static analysis

No suspicious patterns detected.