ClawHub
Back to skill

Security audit

generic-fto-report

Security checks across malware telemetry and agentic risk

Overview

The skill matches its patent FTO purpose, but its external API/AI data flows and MCP setup are inconsistent enough that users should review it carefully before installing.

Install only if you are comfortable sending the relevant product technical descriptions, search strategy, patent claim excerpts, and generated analysis to Zhihuiya/PatSnap services. Resolve the MCP-versus-internal-script setup before use, store the API key securely, and keep generated JSON/report files in an access-controlled location or delete them when no longer needed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill describes capabilities that require reading files, writing reports, and making network/API calls, but it does not declare corresponding permissions or capability boundaries. This creates a transparency and policy-enforcement gap: operators may approve or run the skill without understanding its true access scope, and automated guardrails may not be able to apply least-privilege controls correctly.

Intent-Code Divergence

High
Confidence
95% confidence
Finding
The document gives conflicting trust and execution boundaries: earlier it forbids external MCP/services and says all API calls must stay within built-in scripts, but later it states the skill depends on the Zhihuiya MCP service and agent-enabled MCP tools. This ambiguity can cause the agent or operator to route sensitive patent queries, API credentials, or results through an unintended external integration, weakening control over data flow and auditability.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The document explicitly instructs that AI07 raw output be stored in `fto_structured_data.json` but does not warn about retention, access control, or minimization of potentially sensitive product descriptions, claims text, and model-generated analysis. In an FTO workflow, these artifacts can contain confidential R&D details or legal assessments, so retaining them by default increases exposure if logs, workspaces, or report bundles are later shared or compromised.

Missing User Warnings

High
Confidence
97% confidence
Finding
The function constructs a prompt from document-derived technical features, patent identifiers, titles, and claim text, then sends it to an external AI chat endpoint. In an FTO workflow, those inputs can contain confidential product design details, legal analysis context, or proprietary search strategy; transmitting them to a remote model service without explicit user consent, minimization, or disclosure creates a real confidentiality and data-governance risk.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
references/api_reference.md:15