Security audit

Chat Search

Security checks across malware telemetry and agentic risk

Overview

This skill has a coherent chat-search purpose, but it may search and persist private Feishu or Telegram history without enough scope, permission, or cleanup guidance.

Install only if you are comfortable connecting a local vector database to private chat records. Before use, confirm exactly which Feishu or Telegram data is indexed, how access is authorized, where Qdrant stores it, how to delete it, and require explicit confirmation for vague search requests.

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 (2)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger phrases are broad everyday search terms like “搜索”, “查找”, and “帮我搜”, which can cause this skill to activate for many generic user requests without clear confirmation that chat history should be searched. Because the skill accesses potentially sensitive Feishu/Telegram conversation history, overbroad activation increases the risk of unintended retrieval and disclosure of private messages.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The activation section states the skill should auto-trigger on several vague patterns but does not define boundaries for when it should not run, leaving routing behavior ambiguous. In a skill that searches historical chat data, ambiguity can lead to unauthorized or surprising access to message archives when the user intended a different kind of search.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.