ClawHub
Back to skill

Security audit

pubmed-paper-monitor

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to do what it says: query PubMed for recent journal articles, translate titles, and optionally create a local report, with some privacy and consent caveats.

Install only if you are comfortable with the skill sending PubMed search terms to NCBI, using the embedded Entrez contact email, installing Biopython, and creating a Desktop report for large result sets. For regular use, replace the hardcoded email with a configured contact address and ask before saving reports locally.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill instructs the agent to write a full report directly to the user's Desktop based on article count, introducing local file creation that is not necessary for simply monitoring PubMed papers. Unprompted writes to a prominent user directory can create privacy, clutter, and trust-boundary issues, especially when performed automatically rather than with explicit user consent.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The instructions direct saving a full report to the Desktop without warning the user that a local file will be created or explaining what data will be written. This bypasses informed consent for filesystem side effects and may expose sensitive research interests or create unwanted artifacts on the user's machine.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The script hard-codes a personal email address and automatically sends it to the external NCBI Entrez service on every request. This creates an undisclosed outbound identifier, exposes personal contact information, and can misattribute API usage to a specific individual without user awareness or consent.

Static analysis

No suspicious patterns detected.