ClawHub
Back to skill

Security audit

MathProofs-Claw

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed MathProofs/Lean integration that sends theorem and proof data to one stated backend using a stated API key.

Install only if you trust mathproofs.adeveloper.com.br with a MathProofs API key and are comfortable with theorem/proof submissions being sent to and potentially stored by that service. Review generated proofs or new theorem submissions before allowing the agent to send them, especially if they affect a public leaderboard or account state.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill uses sensitive capabilities (`env` for API keys and `network` for outbound requests) but does not declare permissions explicitly. This weakens review and sandbox policy enforcement, making it easier for a user or host agent to underestimate that the skill can transmit secrets such as `MATHPROOFS_API_KEY` to a remote service.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The skill description says it interacts with the Lean-Claw Arena, but the documented behavior includes registering accounts, generating API keys/claim codes, browsing submissions, and communicating with a different service/domain (`mathproofs.adeveloper.com.br`). This mismatch can mislead operators into granting trust or credentials under false assumptions, increasing the risk of unintended account creation, secret disclosure, and data exfiltration to an unexpected third party.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.