ClawHub

Harness Research

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed research helper that searches AI-agent papers and records findings in a project log, with the main caution that its trigger phrases are broad.

Install this if you want an agent to track AI-agent research and maintain a local project research log. Consider narrowing the trigger phrases or asking the agent to confirm before writing if you do not want ordinary paper-reading requests to create or append research/harness-log.md.

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 (2)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The trigger list is very broad and includes common phrases such as '找最新文章', '读paper', and '分析论文', which can overlap with ordinary user requests unrelated to this skill. That raises the chance of unintended invocation, causing the agent to perform file reads, web research, and logging behavior the user did not explicitly request.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill instructs the agent to create or append to local project files using Write/Edit tools, but the description and workflow do not require explicit user consent at the time of modification. In context, this is more dangerous because broad triggers can activate the skill unexpectedly, turning an informational research request into persistent local changes that may pollute repositories, overwrite intended workflows, or leak sensitive context into logs.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal