Security audit

知识拓展笔记术

Security checks across malware telemetry and agentic risk

Overview

The skill is a note-taking workflow, but it can automatically save user content to external Feishu/Get note destinations and grant full access to a hardcoded Feishu user.

Review this carefully before installing. It is best suited only for the named Feishu workspace/user it was written for. Do not use it for sensitive, proprietary, or personal notes unless you are comfortable with automatic external archiving and a hardcoded Feishu user receiving full access; narrow the triggers and require explicit confirmation before any sync or permission change.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (5)

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The skill explicitly performs a Feishu permission change granting `full_access` to a specific user after document creation. That access-management action is not necessary to analyze knowledge and archive notes, and it expands the blast radius from content creation to authorization changes, enabling unintended sharing or privilege escalation if the skill is triggered on sensitive material.

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The knowledge-base flow repeats the same unjustified `full_access` grant, meaning every archived note may automatically become accessible to a designated account. This creates an unnecessary access-control side effect and increases the risk of unauthorized disclosure or modification of persisted user content.

Vague Triggers

High

Confidence: 94% confidence
Finding: The activation phrases include very common language such as '深度学习', '概念讲解', '新知识', and '我刚学到', making accidental invocation likely during normal conversation. Because activation leads to automatic persistence and external syncing, broad triggers materially increase the chance of unintended data collection, storage, and sharing.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The skill advertises automatic syncing to multiple external destinations plus local file storage, but does not require explicit user acknowledgement of persistence or destination disclosure at execution time. This is dangerous because users may share sensitive notes expecting ephemeral assistance, while the skill silently turns them into durable records across several systems.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The file gives concrete instructions for creating and writing Feishu knowledge-base documents, but it does not warn that user-provided content will be persisted to an external platform. In this skill’s context, users are encouraged to paste newly learned material and notes, so silent syncing to Feishu increases the risk of unintended disclosure of sensitive or proprietary information.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.