ClawHub
Back to skill

Security audit

CLI Design Framework

Security checks across malware telemetry and agentic risk

Overview

This is a documentation-only CLI design aid with no executable code or hidden access; its higher-risk content is illustrative guidance for designing other CLIs.

Install this as design guidance, not as a runnable tool. Treat its example agent-runtime blueprint as a starting point and add your own safeguards for transcript retention, local storage, approvals, and secret handling if you implement those ideas.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The blueprint explicitly stores transcripts and session state locally and says prompts may be logged verbatim, but it does not require a user-facing warning, retention controls, or secret-handling guidance. In a high-side-effect agent runtime, transcripts can contain sensitive prompts, tool outputs, filesystem paths, tokens, and approval history, so silent local persistence creates meaningful confidentiality and compliance risk.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The manifest description is broadly scoped ('designing a new CLI, reviewing an existing CLI, or resolving uncertainty...') and can plausibly activate on a wide range of routine CLI-related prompts without strong boundaries. Over-broad activation can cause the skill to be invoked in contexts beyond its intended use, increasing the chance of irrelevant guidance, prompt-surface expansion, or unintended influence over unrelated tasks.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.