Back to skill

Security audit

High Value Patent Package

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent patent-screening workflow that clearly relies on PatSnap/Zhihuiya APIs and local report files, with no evidence of hidden persistence, destructive behavior, or unrelated data access.

Install only if you are comfortable sending patent search queries and patent identifiers to PatSnap/Zhihuiya and storing the generated report and trace files locally. Prefer environment-variable API keys, avoid shared working directories for confidential searches, and use the Word generator's --noimg option if remote image fetching is not acceptable.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Tainted flow: 'url' from requests.get (line 173, network input) → requests.get (network output)

Medium
Category
Data Flow
Content
def fetch_img(url, timeout=30):
    try:
        r = requests.get(url, timeout=timeout)
        if r.status_code == 200 and r.content[:4] in (b'\x89PNG', b'\xff\xd8\xff\xe0', b'\xff\xd8\xff\xe1', b'GIF8'):
            return io.BytesIO(r.content)
    except Exception:
Confidence
93% confidence
Finding
r = requests.get(url, timeout=timeout)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill clearly instructs network access, local file generation, and use of an environment-provided API token, yet it does not declare these capabilities or permissions. This creates a transparency and least-privilege problem: users or host systems may not realize the skill can exfiltrate query data to external services and persist potentially sensitive outputs locally.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill mandates creation of HTML reports and tracking files containing raw patent evidence and scoring details, but it does not warn the user that files will be written and retained locally. This can expose sensitive search strategies, internal research priorities, or licensed data artifacts to unintended readers on shared systems or downstream workflows.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill requires authenticated API calls to external PatSnap/智慧芽 endpoints using user queries, patent identifiers, and related metadata, but it does not disclose that this information will be transmitted off-box to third-party services. In environments where queries reflect confidential R&D, litigation, or acquisition strategy, undisclosed transmission can create privacy, contractual, or data-governance risk.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The code silently downloads remote images while generating a document, which creates implicit network access and leaks that report generation occurred to third-party hosts. In this skill context, the image URLs are data-driven and may point to untrusted locations, so the lack of disclosure also obscures a real security/privacy boundary from operators.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.