Back to skill

Security audit

SafeRoute Base swap preflight

Security checks for vulnerabilities and agentic risk

Overview

The skill provides a disclosed swap-check service with an unpaid demo and payments controlled by the operator’s existing client.

Start with the unpaid demo. For paid use, require your existing client to enforce the approved merchant, Base USDC, and a 0.001 USDC limit; quotes need a separate 0.01 USDC allowance. Keep wallet secrets local and attribution labels non-personal. The live service and external payment client 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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, network.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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 23)May include surrounding context.

Evaluate with the operator’s existing payment client

Use POST https://api.usesaferoute.com/v1/swap-safety-check with:

json
{ "from": "USDC", "to": "WETH", "amount": "1" }

External Transmission

Medium
Category
Data Exfiltration
Confidence
50% 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 33)May include surrounding context.

md
The bundled helper exports `runFirstCall({fetchImpl, paidFetch, source, applicationId})` for an existing caller-controlled x402 fetch. Passing paidFetch permits one paid service request after the unpaid inspection; callers must apply the above payment policy. Do not add automatic retries after uncertain settlement. A successful payment response plus a received service result establishes that call only.

For an optional quote use `POST /v1/quote`, or MCP tool `get_safe_swap_quote` at `https://api.usesaferoute.com/mcp`, after validating the live schema and obtaining the separate 0.01 USDC allowance. MCP safety tool: `check_base_swap`. Do not escalate to signing, broadcasting or a funded trade from a request to evaluate an API.

## Learn what worked

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 29)May include surrounding context.

{ "from": "USDC", "to": "WETH", "amount": "1" }

text

Use the live [OpenAPI](https://api.usesaferoute.com/openapi.json) for supported assets and current schemas. Inspect the challenge first. Before a paid call, verify scheme exact, Base `eip155:8453`, canonical USDC `0x833589fCD6eDb6E08f4C7C32D4f71b54bdA02913`, the intended merchant and an application-controlled spend ceiling. Keep wallet credentials in the caller’s existing local payment client; never send them to SafeRoute or ask for seed phrases.

The bundled helper exports `runFirstCall({fetchImpl, paidFetch, source, applicationId})` for an existing caller-controlled x402 fetch. Passing paidFetch permits one paid service request after the unpaid inspection; callers must apply the above payment policy. Do not add automatic retries after uncertain settlement. A successful payment response plus a received service result establishes that call only.

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 · scripts/inspect.mjs (reported line 3)May include surrounding context.

js
import { pathToFileURL } from "node:url";

export const endpoint = "https://api.usesaferoute.com/v1/swap-safety-check";
export const input = { from: "USDC", to: "WETH", amount: "1" };
export const payTo = "0x08a6a2ddecdebdbfcb29d725042a6f3d74849e08";
export const usdc = "0x833589fcd6edb6e08f4c7c32d4f71b54bda02913";