ClawHub
Back to skill

Security audit

Notion技能

Security checks across malware telemetry and agentic risk

Overview

This Notion skill mostly matches its stated purpose, but it also asks for broad command execution and local file authority that are not clearly bounded to Notion tasks.

Install only if you are comfortable giving the agent Notion workspace mutation authority and broad local command capability. Use a least-privilege Notion integration shared only with the pages or databases needed, keep the token in an environment variable, and explicitly confirm any create, update, or delete action before it runs.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The skill presents command execution as a user-facing capability despite being a Notion API integration, which expands the attack surface far beyond the stated purpose. In an agent environment, normalizing exec without strict scoping can enable prompt-driven shell execution, unsafe local actions, or abuse of host access unrelated to Notion operations.

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
Declaring exec capability for a skill whose core purpose is interacting with the official Notion API grants unnecessary privileges. Excess capability is dangerous because an attacker can coerce the agent to perform local command execution or system access that is unrelated to the Notion workflow, increasing the blast radius of prompt injection or misuse.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The skill documents create, update, and delete operations against workspace data without prominently warning users that these actions are destructive or modifying. In an agent setting, unclear disclosure raises the risk of accidental data loss, unauthorized modifications, or users invoking dangerous actions without understanding the consequences.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.