Back to skill

Security audit

AgentForge DeFi Safety

Security checks for vulnerabilities and agentic risk

Overview

The skill offers clearly disclosed cryptocurrency safety checks, with privacy and optional payment considerations.

Start with the wallet-free MCP option. Submitted wallet addresses and planned transaction or signature details reach AgentForge, so confirm that sharing is acceptable and never submit private keys or seed phrases. For paid use, constrain the payment wallet and set a spending limit before enabling polling or follow-up checks. Treat safety verdicts as advice, not guarantees; the deployed backend and retention practices were not verified.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

The preflight workflow sends planned transactions and signature requests to an external service without a privacy warning.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The blanket instruction to submit upcoming transactions and signature requests to a third-party preflight service can disclose nonpublic trading intentions, counterparties, and wallet associations before execution. This transmission serves the skill's stated safety purpose, but the workflow provides no privacy warning or consent boundary for sensitive pre-broadcast data; no malicious collection or theft of signing keys is demonstrated.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 102)May include surrounding context.

1. Wallet safety check

bash
# Step 1: unpaid request — returns 402 with payment terms
curl -i -X POST https://agentforge-taupe.vercel.app/v1/wallet-safety \
  -H "Content-Type: application/json" \
  -d '{"walletAddress":"0x7Fc66500c84A76Ad7e9c93437bFc5Ac33E2DDaE9","chain":"ethereum"}'

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 134)May include surrounding context.

js
import { wrapFetchWithPayment } from '@x402/fetch';
const paidFetch = wrapFetchWithPayment(fetch, x402Client); // handles the 402 -> X-PAYMENT retry for you
const res = await paidFetch('https://agentforge-taupe.vercel.app/v1/wallet-safety', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({ walletAddress: '0x...', chain: 'ethereum' }),