ClawHub
Back to skill

Security audit

Harness Dev Standards

Security checks across malware telemetry and agentic risk

Overview

This is a coherent development-quality skill, but it tells agents to auto-fix code, dependencies, environment files, and port conflicts without enough user control.

Review this skill carefully before installing. Use its checklists and scripts as advisory tools, but require explicit approval before dependency changes, source edits, npm audit fix, global package installation, .env edits, or terminating processes. Check diffs and keep secret values out of commits and logs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill advertises very broad activation criteria such as starting new projects, pre-delivery checks, standardized development flow, architecture review, code review, and troubleshooting dependencies or environments. This can cause the agent to invoke the skill in many routine development contexts without clear scoping, increasing the chance of unnecessary or inappropriate guidance being applied across unrelated repositories or tasks.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The auto-remediation section instructs automatic fixes for dependencies, imports, syntax, startup failures, and type errors, with a repeat-until-resolved workflow, but it does not warn that project files, lockfiles, or dependency versions may be changed. In an agent setting, this encourages potentially destructive autonomous modifications that can introduce supply-chain risk, break builds, or alter project behavior without explicit approval or review.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The guide recommends using `kill -9 <PID>` to clear a port conflict but does not warn that SIGKILL forcibly terminates the target process without cleanup, which can cause data loss, corruption, or interruption of unrelated services if the wrong PID is selected. In an automated remediation context, destructive commands without guardrails are dangerous because users or agents may apply them mechanically.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The guide instructs users to create or update a `.env` file with missing environment variables but does not warn that these values often contain secrets such as database credentials, API keys, or tokens. In a remediation skill, this can normalize unsafe secret handling, increasing the risk of accidental exposure through commits, logs, screenshots, or overly permissive file access.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.