ClawHub
Back to skill

Security audit

OpenSpec

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent OpenSpec workflow helper, but users should review its global npm install and repository-changing commands before use.

Install this only if you trust the OpenSpec npm package, and consider pinning a version instead of using `@latest` for reproducible or sensitive projects. Use version control, inspect generated `openspec/` and `.claude/skills/` files, and review changes before running archive or accepting implementation edits.

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
88% confidence
Finding
The trigger description is very broad and can cause the skill to activate on routine development prompts such as feature work, refactors, planning, or any mention of /opsx commands. In an agent setting, overbroad auto-invocation increases the chance the agent follows this workflow unnecessarily and begins making structured repository changes or running CLI commands without the user clearly requesting OpenSpec-specific behavior.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The documented `openspec archive <name> --yes` command suppresses confirmation for an operation that merges specs and moves change artifacts, which can alter repository state in a non-trivial way. Presenting this as a standard step without warning makes it more likely an agent will execute a destructive or irreversible action automatically, especially in unattended or tool-enabled environments.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The agent workflow instructs the agent to create files under `openspec/changes/<name>/`, implement tasks, validate, and archive, but it does not require explicit user consent before modifying the repository. In a tool-using agent, these instructions can translate directly into file writes and broader code changes, creating a risk of unintended modification beyond what the user expected.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal