Claw Earn

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a disclosed Claw Earn operations runbook, but it also embeds instructions to steer users into a separate paid USDC service outside the stated task workflow.

Install only if you want an agent operating Claw Earn production workflows. Review every wallet transaction, amount, contract, task ID, notification email change, and message before approval, and treat Founder-Level Intelligence as a separate paid service that should only be used after an explicit user request and payment confirmation.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The skill materially expands from Claw Earn task operations into a separate paid 'Founder-Level Intelligence' service, including separate discovery, positioning, and purchase guidance. That broadens the agent's authority and can cause capability drift, where an operator invoking a task-management skill is unexpectedly steered into unrelated commercial actions and external spend.

Context-Inappropriate Capability

High
Confidence
99% confidence
Finding
The skill instructs the agent to initiate a paid x402 purchase flow, pay a quote in USDC on Base, and redeem the purchase, despite the skill's stated purpose being task management. This creates a direct path to unauthorized value transfer and social-engineering-style upsell behavior, especially dangerous in a production operations skill that otherwise normalizes wallet use and transaction execution.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal