ClawHub
Back to skill

Security audit

arXiv 论文推荐器

Security checks across malware telemetry and agentic risk

Overview

This skill searches public arXiv and GitHub data for Agent/RAG papers and saves local recommendation reports as described.

Install if you are comfortable with a skill that contacts arXiv and GitHub and keeps recommendation reports plus history under ~/papers. On shared machines, review or delete ~/papers/history.json and ~/papers/recommendations if you do not want that local history retained.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (3)

Description-Behavior Mismatch

Medium
Confidence
92% confidence
Finding
The script writes persistent data under the user's home directory for both recommendation history and generated reports, but this storage behavior is not disclosed in the skill description. Undisclosed filesystem writes create privacy and transparency risks because paper titles, topics, and timestamps may be stored without the user's awareness and can accumulate over time.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger phrases are broad enough to match ordinary discussion about papers, arXiv, or recent research, so the skill could activate when the user did not intend to run a workflow involving network access and local file writes. Unintended activation increases the chance of surprise external requests, disk writes, and persistence of recommendation history without meaningful user intent.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill writes a generated report to ~/papers/recommendations and stores recommendation history in ~/papers/history.json, but the description does not clearly warn users about this persistence behavior before use. Silent local storage can expose sensitive research interests, overwrite existing files, or accumulate data beyond user expectations, especially on shared systems.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.