ClawHub
Back to skill

Security audit

Craft

Security checks across malware telemetry and agentic risk

Overview

This Craft skill is mostly a documented integration helper, but it recommends destructive production-document replacement in a way users should review carefully before installing.

Install only if you are comfortable with an agent receiving instructions for broad Craft production writes. Before using it on existing documents, require explicit confirmation for deletes, moves, full rewrites, task or collection changes, and keep a backup or export for content that matters.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The skill is marketed as a document creation/formatting helper, but its documented tool surface includes broad write capabilities such as delete, move, task, collection, whiteboard, and comment operations. This scope mismatch increases the chance that an agent will invoke destructive or unrelated actions under the guise of routine document editing, violating least privilege and user expectations.

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The rewrite guidance explicitly recommends deleting and recreating whole documents on a production system as a 'fast' path. That is dangerous because it can cause irreversible data loss, break links/metadata/history, and normalize destructive behavior for a skill whose stated purpose is formatting and editing.

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill warns that it operates on production data and labels permanent deletions as unsafe, yet later recommends whole-document deletion/recreation. This contradiction undermines operator safety, making risky actions seem acceptable and increasing the likelihood of accidental destructive changes in a real user workspace.

Vague Triggers

High
Confidence
88% confidence
Finding
The trigger language says the skill activates on essentially any Craft document or note creation, editing, or formatting task, which is overly broad. Broad activation increases the chance the skill is selected in situations where its powerful write capabilities are unnecessary, leading to unintended mutations on production content.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.