ClawHub
Back to skill

Security audit

Tasks

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a task-management helper, but its credential and backend documentation is broader and less scoped than a Todoist-only tool should be.

Review the skill carefully before installing. Use it only if you expect it to access and modify your task account, prefer a narrowly scoped Todoist token, and do not provide a Microsoft Graph token unless the publisher documents exactly which Graph permissions are required and why.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The skill documentation claims support for Microsoft To-Do and use of MSGRAPH_TOKEN even though the manifest and examples are centered on a Todoist CLI. This scope expansion can mislead users about what software, credentials, and API surfaces are involved, increasing the chance of improper token exposure, unsupported execution paths, or accidental use of a different service than intended.

Intent-Code Divergence

Low
Confidence
79% confidence
Finding
The file presents the skill as a CLI wrapper but the installation instructions reference a Python API package and token-based API usage without clearly defining the trusted executable or command behavior. This mismatch can cause users to install the wrong component, rely on undocumented code paths, or grant credentials to tooling whose scope is unclear.

Missing User Warnings

Low
Confidence
75% confidence
Finding
The documentation instructs use of TODOIST_API_TOKEN or MSGRAPH_TOKEN without warning users about secure storage, shell history leakage, process inspection, or accidental disclosure in logs. In a credential-driven skill, omission of basic handling guidance increases the likelihood of secrets being exposed during setup or troubleshooting.

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.