ClawHub

Ai测试 09ed64

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only Chinese AI testing guide with broad trigger wording and some promotional or off-topic copied content, but it does not request dangerous access or perform actions.

Install only if you want a lightweight Chinese note-style guide about AI testing and related API relay references. Consider narrowing or removing the generic "AI" trigger so it does not appear during unrelated AI conversations, and review the repeated promotional Bilibili/API relay content before relying on it.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

High
Confidence
97% confidence
Finding
Using the trigger phrase "AI" makes activation far too broad, so the skill may be invoked during many unrelated conversations. That creates an unsafe routing condition where irrelevant or lower-trust skill content can hijack user intent, increasing the chance of prompt injection, confusing responses, or accidental disclosure through unintended execution.

Vague Triggers

High
Confidence
98% confidence
Finding
The trigger scenario explicitly activates on the phrase 用户说"AI", which is so generic that it will match a huge fraction of normal user requests. In an agent environment, this broad matching can cause untrusted skill content to preempt more appropriate tools or inject irrelevant instructions into unrelated workflows.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The condition 用户需要AI测试相关帮助 is ambiguous and lacks concrete boundaries, leaving activation to subjective interpretation by the routing layer. Ambiguous scope increases the likelihood of accidental triggering and makes it easier for loosely related prompts to pull in this skill unexpectedly.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal