ClawHub
Back to skill

Security audit

医学知识图谱mindmap

Security checks across malware telemetry and agentic risk

Overview

This is a local medical knowledge-graph skill that stores user-entered facts on disk; the storage is purpose-aligned and disclosed, but users should avoid putting private patient data into it.

Install only if you want a local persistent study knowledge base. Be aware that facts and relationships you add are saved in local JSON/JSONL files, and the artifacts do not show encryption, retention controls, or deletion tooling; avoid storing patient-identifying health information or confidential clinical records.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill documents file read/write behavior by specifying persistent storage paths under memory/ and giving command examples that create and update graph.jsonl and knowledge files, but it declares no permissions. Undeclared filesystem access weakens user/admin consent and policy enforcement, and broadens the blast radius if the skill is invoked unexpectedly or manipulated through its natural-language triggers.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger phrases are very broad conversational utterances such as asking relationships, summaries, or whether a waveform is epilepsy-related, without strict invocation boundaries. Because the skill can perform persistent writes and relationship creation, overly broad activation increases the risk of accidental invocation, unintended data modification, or unreviewed medical-content generation from ordinary conversation.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The script persists user-supplied medical knowledge facts to local JSON files without any notice, consent, retention policy, or access controls. In a medical context, facts may contain patient-linked or otherwise sensitive health information, so silent persistence increases privacy and compliance risk if users assume inputs are ephemeral.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The initialization and graph-writing logic creates persistent local storage for entity and relationship data without clear user disclosure. Because this skill manages medical knowledge graphs, stored graph nodes and relationships may reveal sensitive diagnostic, treatment, or case-associated information if used with real-world data.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.