ClawHub
Back to skill

Security audit

Novel Generator Skip

Security checks across malware telemetry and agentic risk

Overview

This is a coherent novel-writing skill that stores story files locally and has a scoped cleanup option users should handle carefully.

Install it only in a workspace where local novel drafts and continuity notes can be saved. Back up output/ and .learnings/ before using --clean, avoid putting sensitive personal information into prompts or story memory, and confirm any optional image-generation tool before sending story content outside the local workspace.

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 (3)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill’s trigger description is broad enough to match many generic writing requests, which can cause the skill to activate outside the user’s actual intent. In an agent environment with file read/write permissions, over-triggering increases the chance of unnecessary filesystem access and unexpected autonomous actions such as creating or modifying project files.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The documentation instructs use of an initialization script that clears prior `.learnings/` records but does not require an explicit warning, preview of affected files, or confirmation step. This creates a real data-loss risk because users may invoke the script expecting setup behavior without realizing existing story memory and prior work will be erased.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The --clean option performs destructive deletion of prior output without any confirmation prompt or dry-run safeguard. In an agent-driven workflow, a mistaken or injected invocation using --clean can silently erase existing novel content, causing data loss even though the deletion is scoped to the skill's output directory.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.