ClawHub
Back to skill

Security audit

llm-provider AI工具

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed OpenAI/llm-provider management helper, with external API and remote resource actions that users should approve carefully.

Install only if you intend to let the agent use ClawLink/OpenAI-style tools. Review any preview before writes, uploads, assistant/vector-store creation, batch submission, or fine-tuning, and avoid sending sensitive local files unless you are comfortable with them being processed by the external provider.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The skill description is extremely broad, advertising generic AI model usage, agent orchestration, workflow automation, and multiple OpenAI resource-management functions. This can cause the skill to be selected for loosely related user requests, increasing the chance of unintended external API calls, file handling, or command execution beyond what the user explicitly expected.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill documents file operations, uploads, model calls, assistant creation, vector store creation, and batch actions, but does not present a prominent upfront warning that data may be transmitted to external services and that actions may create or modify remote resources. In this context, the risk is elevated because the declared tools include read, write, and exec, and examples include uploading local files and invoking remote APIs, which can lead to accidental data disclosure or unintended side effects.

Static analysis

No suspicious patterns detected.