ClawHub
Back to skill

Security audit

Pubmed Verifier

Security checks across malware telemetry and agentic risk

Overview

This is a coherent PubMed citation-checking skill with normal network, cache, and report behavior for that purpose, though users should be mindful when scanning private drafts.

Install this if you need biomedical citation or PMID verification. Before running it on unpublished manuscripts, internal notes, or sensitive projects, review the chosen --source path, consider --no-cache if you do not want PubMed metadata stored locally, and remember that --suggest can send citation-derived search terms to PubMed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger list is very broad and includes common phrases like 'check citations', 'validate references', and Chinese equivalents that can appear in ordinary research or writing workflows. That raises the chance of unintended invocation, which matters here because the skill can scan project files, send citation-derived data to external APIs, and generate local reports or cache entries without the user fully realizing the side effects.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill description emphasizes functionality but does not clearly disclose that citation context and metadata extracted from local files may be transmitted to PubMed and Crossref, nor that results are cached in SQLite and reports may be written locally. This creates a privacy and data-handling risk, especially when source files contain unpublished manuscripts, internal notes, or sensitive project content embedded near citations.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.