Back to skill

Security audit

Shop Car Insurance

Security checks across malware telemetry and agentic risk

Overview

This is a coherent CreditClaw shopping and payments skill, but it gives an agent real spending and payment abilities that need tight owner controls.

Install only if you intentionally want an agent to buy goods or services, request top-ups, create payment links, or use x402 payments through CreditClaw. Keep ask-for-everything approval enabled at first, set low per-transaction and daily limits, restrict merchants/categories where possible, protect CREDITCLAW_API_KEY in a secrets manager, avoid logging request bodies or headers, and review auto-approval, payment-link, heartbeat, and x402 settings before enabling them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (15)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill's manifest frames the capability as guarded agent shopping, but the documented API also lets the agent generate payment links and collect money from third parties. This materially expands the operational scope from spending to payments/merchanting, which can surprise operators, increase abuse surface, and enable unauthorized billing workflows if the agent is prompted to monetize activity.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The manifest emphasizes owner-approved shopping, but the skill also supports x402 signing and agent-to-agent payments, which are broader financial operations than ordinary purchasing. Hidden expansion into signing/payment protocols increases the chance that an operator enables the skill expecting retail purchases only, while the agent gains the ability to initiate transfers in other ecosystems.

Description-Behavior Mismatch

Low
Confidence
84% confidence
Finding
The API reference exposes a 'Sub-Agent Card' rail that is not described elsewhere in the manifest, indicating undocumented capability surface. Undocumented financial rails are risky because defenders and operators cannot accurately assess what the skill may permit or monitor, even if the endpoint is only informational in this file.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The documentation instructs agents to submit a full shipping address, product details, and purchase metadata to an external payment and ordering service without clearly warning about the privacy implications. This can lead to unnecessary disclosure of personally identifiable information and purchase behavior to third parties, especially if users or integrators assume the action is purely local or internal.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to automatically send a top-up request when balance is low, which is a state-changing external action on a financial system. Although it later informs the human, the user is not clearly warned beforehand or asked for consent before the request is sent, creating a risk of unauthorized financial workflow initiation.

Missing User Warnings

Medium
Confidence
84% confidence
Finding
The documentation instructs agents to transmit sensitive personal and commercial data, including shipping address, product details, and authorization credentials, to an external purchasing API without any explicit privacy, consent, retention, or data-sharing warning. In an agent skill context, this is risky because it normalizes exfiltration of owner/user PII and purchase metadata to third-party infrastructure and merchants, even if the business purpose is legitimate.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
This documentation instructs an agent to initiate real-world purchases and highlights auto-approval behavior, but it does not prominently warn that these actions can spend real funds. In an agent-skill context, omission of an explicit spending-risk warning increases the chance that downstream integrators or autonomous agents will treat the flow as routine API usage and trigger unintended transactions.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The guide instructs an agent to initiate real USDC payments and describes on-chain settlement, but it does not prominently warn that these actions can spend real funds, may be irreversible, and depend on externally supplied payment details. In an agent skill context, omission of a clear financial-risk warning can lead to unsafe autonomous payment behavior or operator misunderstanding.

Missing User Warnings

Low
Confidence
88% confidence
Finding
The example shows use of a bearer API key in Authorization headers without any warning that the credential is sensitive and must not be logged, shared, or embedded in client-side contexts. In agent and automation setups, such omissions increase the chance of accidental key exposure through transcripts, debugging output, or copied examples.

External Transmission

Medium
Category
Data Exfiltration
Content
### Purchase Request

```bash
curl -X POST https://creditclaw.com/api/v1/card-wallet/bot/purchase \
  -H "Authorization: Bearer $CREDITCLAW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
92% confidence
Finding
curl -X POST https://creditclaw.com/api/v1/card-wallet/bot/purchase \ -H "Authorization: Bearer $CREDITCLAW_API_KEY" \ -H "Content-Type: application/json" \ -d

External Transmission

Medium
Category
Data Exfiltration
Content
### Purchase Request

```bash
curl -X POST https://creditclaw.com/api/v1/card-wallet/bot/purchase \
  -H "Authorization: Bearer $CREDITCLAW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
81% confidence
Finding
curl -X POST https://creditclaw.com/api/v1/card-wallet/bot/purchase \ -H "Authorization: Bearer $CREDITCLAW_API_KEY" \ -H "Content-Type: application/json" \ -d

External Transmission

Medium
Category
Data Exfiltration
Content
### Checkout Request

```bash
curl -X POST https://creditclaw.com/api/v1/bot/merchant/checkout \
  -H "Authorization: Bearer $CREDITCLAW_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
89% confidence
Finding
curl -X POST https://creditclaw.com/api/v1/bot/merchant/checkout \ -H "Authorization: Bearer $CREDITCLAW_API_KEY" \ -H "Content-Type: application/json" \ -d

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## How It Works

Self-hosted cards use a split-knowledge privacy model. Your owner provides their own card details through CreditClaw's secure setup wizard — you never see the actual card numbers. When you need to make a purchase at any online merchant, you submit a checkout request. CreditClaw evaluates it against your card's permissions and either auto-approves (if within your allowance) or sends your owner an approval request via email.

**Use this rail for:** Any online store — SaaS subscriptions, cloud hosting, domain registrations, digital services, or any merchant not covered by the Pre-paid Wallet.
Confidence
95% confidence
Finding
auto-approve

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
1. You submit a checkout request with merchant and amount details
2. CreditClaw evaluates the request against your card's permissions
3. If the amount is within your auto-approved allowance, it processes immediately
4. If the amount exceeds the threshold, your owner receives an approval request (email with secure link)
5. You poll for the result
6. Once approved, the transaction is recorded
Confidence
96% confidence
Finding
auto-approve

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## Allowance Thresholds

Your owner sets a per-profile allowance threshold for each card. Purchases within this threshold are auto-approved — no email confirmation needed. Purchases above it require human approval via a secure email link (15-minute TTL).

Your owner can view and adjust these thresholds from their dashboard at `https://creditclaw.com/app/self-hosted`.
Confidence
94% confidence
Finding
auto-approve

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.