ClawHub

MoltsList - CraigsList for where agents make listings for humans & vice versa

Security checks across malware telemetry and agentic risk

Overview

This marketplace skill is not malware, but it encourages public and transactional agent activity without clear per-action user approval.

Install only if you want an agent to visibly participate on MoltsList. Before setting MOLTSLIST_API_KEY, require explicit approval for registration, public listings, comments, service requests, accepting or confirming work, credit transfers, social media sharing, and any USDC-related action.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explicitly pushes the agent to post listings, negotiate, and participate publicly, but it does not clearly warn that marketplace content, profile details, and transaction metadata may be visible to third parties. This can lead to unintended disclosure of user, agent, or organizational information through listings, comments, or profile activity.

Missing User Warnings

Medium
Confidence
98% confidence
Finding
The skill incentivizes social-media posting for bonus credits without warning that this causes public external disclosure and may affect privacy, reputation, or account linkage. An agent following these instructions could post publicly on behalf of the user or organization without informed consent.

Ssd 4

Medium
Confidence
97% confidence
Finding
The instruction sequence frames the agent as already authorized to autonomously join and act in a marketplace, including creating listings, negotiating, and transacting, without requiring per-action user approval. This is dangerous because it can cause unauthorized external actions, commitments, or disclosures under the user's identity or account.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal