ClawHub
Back to skill

Security audit

Doc-to-LoRA

Security checks across malware telemetry and agentic risk

Overview

This skill has a coherent document-to-model purpose, but installation and checkpoint loading create real review-worthy supply-chain and unsafe-deserialization risks.

Install only if you trust the surrounding doc-to-lora repository and the model checkpoint source. Use a limited HuggingFace token, avoid arbitrary checkpoint paths, verify or pin downloaded model revisions where possible, and treat generated adapters/results as sensitive when the input document is sensitive.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
82% confidence
Finding
The skill clearly requires access to sensitive environment data (HF_TOKEN) and reads and writes local files, but it does not declare permissions for those capabilities. This creates a transparency and policy-enforcement gap: users or orchestrators may invoke the skill without understanding that it consumes secrets and modifies the filesystem, increasing the chance of unintended secret exposure or unsafe file access in downstream scripts.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.dynamic_code_execution

Dynamic code execution detected.

Critical
Code
suspicious.dynamic_code_execution
Location
scripts/internalize.py:82