AMEX | Give your Agent your CreditCard

WarnAudited by ClawScan on May 18, 2026.

Overview

This is a disclosed payment-wallet skill, but it asks an agent to handle real spending authority, decrypted card data, and a remotely delivered decrypt script, so it needs careful review before use.

Install only if you trust CreditClaw and have verified the publisher. Keep the default ask-before-everything mode, use low-limit or virtual cards, avoid the main-agent card-decryption fallback, sandbox any decrypt-script execution, protect the CREDITCLAW_API_KEY, and monitor or freeze the wallet from the dashboard if anything looks wrong.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

A user may think this is associated with a known card issuer when the artifacts show a different provider.

Why it was flagged

The registry-facing name invokes AMEX, but the homepage and included files describe a CreditClaw service rather than an AMEX/American Express service. For a payment-card skill, that branding mismatch could mislead users about affiliation or trust.

Skill content
Name: AMEX | Give your Agent your CreditCard ... Source: unknown ... Homepage: https://creditclaw.com
Recommendation

Verify the publisher and any claimed card-network or issuer affiliation before providing payment credentials or funds.

What this means

Anyone or any agent that obtains the key may be able to spend or act through the wallet within the provider's guardrails.

Why it was flagged

The required CREDITCLAW_API_KEY is expected for this payment service, but the artifact itself confirms that the key represents delegated spending authority.

Skill content
Your API key is your identity. Leaking it means someone else can spend your owner's money.
Recommendation

Keep the key scoped to this service, store it securely, rotate it if exposed, and use low spending limits and approval-required defaults.

What this means

If the delivered script or delivery channel is compromised, local code could run in the checkout environment with access to sensitive card data.

Why it was flagged

The skill instructs an agent/sub-agent to execute a Node script delivered later inside a card file, not a script included in the reviewed package, while handling real payment-card material.

Skill content
The sub-agent runs the deterministic decrypt script that was delivered with the card file: node decrypt.js <key_hex> <iv_hex> <tag_hex> Card-ChaseD-9547.md
Recommendation

Only use this in a sandboxed environment, inspect or pin the decrypt script before execution where possible, and avoid running the flow on systems with unrelated secrets.

What this means

Your card number, CVV, expiry, name, and billing address could become visible to the main agent context rather than isolated to an ephemeral checkout agent.

Why it was flagged

The artifact acknowledges a fallback path where full card details enter the main agent context, which may be logged, retained, or reused outside the intended single transaction depending on the host environment.

Skill content
If your environment doesn't support spawning sub-agents, you can execute the checkout_steps directly as the main agent ... the main agent will see the decrypted card details in its context.
Recommendation

Do not use the main-agent fallback for real cards unless you understand the host's logging and memory retention; prefer disposable virtual cards, sub-agent isolation, and strict cleanup.

What this means

The agent may periodically check balances and spending permissions if you configure it to do so.

Why it was flagged

The periodic heartbeat is purpose-aligned wallet monitoring and no daemon install is shown, but it encourages recurring autonomous use of the payment API key.

Skill content
CreditClaw Heartbeat (suggested: every 30 minutes) ... Run this routine periodically to stay aware of your wallet status
Recommendation

Allow periodic polling only if you want ongoing wallet monitoring, and keep rate limits and notification settings enabled.