ClawHub

archtree-community-operator

Security checks across malware telemetry and agentic risk

Overview

This instruction-only skill is transparent about using an Archtree account to read and write community content, with some broad trigger wording that users should scope carefully.

Install this only if you want an agent to operate your Archtree community account. Keep the bearer token private, be explicit about read-only versus write permission, and require fresh confirmation before public posts, replies, edits, deletes, or ongoing proactive participation.

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 description is extremely broad and includes generic terms like community, channels, posts, reply, like, and recent discussions, which can cause the skill to activate for loosely related requests. In a live-instance skill with read/write capabilities, unintended invocation raises the risk of operating on the wrong site context or initiating sensitive community actions when the user did not actually intend to use this skill.

Vague Triggers

Medium
Confidence
88% confidence
Finding
The file defines proactive-mode triggers using broad phrases like checking what people have been discussing lately or staying engaged on an ongoing basis, which can cause the agent to enter a write-capable patrol workflow from loosely worded user requests. In a skill that supports reading, posting, replying, and liking in a live community, this ambiguity increases the risk of overbroad autonomous action and unintended posting beyond what the user specifically requested.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal