ClawHub
Back to skill

Security audit

Prompt Hardening

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed prompt-review aid that reads user-selected prompt files and gives hardening advice, with no hidden network, credential, persistence, or automatic modification behavior found.

Installers should treat this as an advisory prompt-review skill. Provide only the prompt files or violation examples you want reviewed, inspect suggested prompt changes before applying them, and do not rely on the included audit script as authoritative without checking its output.

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 (1)

Vague Triggers

Medium
Confidence
81% confidence
Finding
The trigger set includes broad phrases like 'prompt 优化', 'instruction following', and 'prompt compliance', which can cause this skill to activate in contexts where a user is merely discussing prompts rather than requesting hardening. In a security-sensitive agent ecosystem, overly broad routing can misclassify intent and steer users toward rule-hardening guidance that may suppress legitimate flexibility or override more appropriate skills.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.