Observed arrival · 2026-11-11
Attack My Agent puts AI payment workflows through a hostile test range
Consistra’s testnet tool runs an AI agent through attacks involving payment requests, invoices, approvals, and secrets, then produces a scored report of its actions.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams testing agents with payment tools
- Worth noticing
- Runs use fresh addresses and amounts; the page says no transactions are signed, forwarded, or settled.
Field notes
The hosted test uses an agent’s usual browsing and payment tools against a testnet wallet, recording payments, approvals, sign-ins, and attempted secret disclosures. The page says each run uses fresh addresses and amounts, and that reports remain available by link for 30 days. It also describes a command-line runner and Python package as forthcoming; for now, users are directed to the hosted URL.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
⊠LoginAccess appeared gated
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue