ClawHub
Back to skill

Security audit

SJTU SLURM Skill

Security checks across malware telemetry and agentic risk

Overview

This SJTU HPC helper is sensitive because it stores tokens and SSH credentials, but the artifacts disclose that behavior and keep it aligned with managing the user's HPC account, jobs, and data.

Install only in a private, trusted workspace. Do not paste your HPC password into chat if avoidable; use the temporary-file flow and remove or revoke stored tokens, SSH keys, and certificates when no longer needed. Confirm any job cancellation, account change, data deletion, or large transfer before allowing it to proceed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill performs sensitive operations including file read/write, shell execution, and network access, yet declares no explicit permissions or constraints. This creates a transparency and governance gap: an agent could handle credentials, invoke remote commands, and access external systems without the platform or user having a clear permission boundary.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The description says the skill logs into the HPC platform for job and data operations, but the body also requests and refreshes tokens, generates SSH credentials, stores private keys and tokens locally, and handles password-bearing files. That mismatch is security-significant because users and reviewers may underestimate the credential-management and persistence capabilities granted to the skill.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.