ClawHub
Back to skill

Security audit

opensource-llm-models

Security checks for vulnerabilities and agentic risk

Overview

This skill is a model-selection and deployment guide with expected API examples, and I found no hidden execution, persistence, credential collection, or destructive behavior.

Before copying the API examples, treat prompts and metadata as data sent to the named third-party provider; avoid secrets, regulated data, or proprietary content unless that provider is approved for your use case. Local deployment examples are preferable for sensitive data.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill provides API examples that send user prompts to third-party model providers but never warns about privacy, retention, jurisdiction, or data-sharing implications. In a skill focused on model selection and API usage, users may copy these examples directly and unknowingly transmit sensitive prompts or proprietary data off-platform.

External Transmission

Medium
Category
Data Exfiltration
Content
import openai
client = openai.OpenAI(
    api_key="sk-xxx",
    base_url="https://api.deepseek.com/v1",
)
response = client.chat.completions.create(
    model="deepseek-chat",
Confidence
87% confidence
Finding
The example explicitly configures an external API endpoint, meaning prompts are transmitted outside the local environment to a third-party service. In isolation this is expected for API usage documentation, but it becomes security-relevant because the skill does not pair the example with warnings about sensitive data exposure, provider trust boundaries, or retention policies.

Static analysis

No suspicious patterns detected.