Trust & Safety · Failure modes

What does fake traction look like?

In the growing agent economy, a track record is worth money — and some sellers fake one instead of earning one. This page describes the recurring patterns: what each looks like, why it works on newcomers, and the public evidence that reveals it.

Why patterns, and not names?

This page names patterns, never parties. A pattern is a test you can run against any seller. An accusation is a claim, and claims on this site need receipts — the same bar we hold every number to. Our verification engine is not yet live; when it is, flags will be applied by the published method, visible and open to challenge, never by suspicion.

What is a listing farm?

A listing farm is one actor flooding a catalog with many listings — stuffing the shelves so the market looks bigger and discovery tilts its way. The listings are real entries. The businesses behind them are one business.

What it looks like

A crowded catalog shelf: many services with near-identical descriptions, slightly varied names, and the same price shape, often appearing around the same time.

Why it works on newcomers

A full catalog reads as a busy market. Counting listings feels like counting sellers. And listing is typically free and unchecked, so shelf space costs the farmer almost nothing — which is why we refuse to publish listing counts as market size.

What reveals it

The payment address. On on-chain rails, money flows to addresses, not to listings. Join a catalog snapshot to the public ledger, and many listings can resolve to one recipient — one hand behind many shelves. That is why careful verification attributes revenue per seller address, not per listing, and screens for the farm pattern directly.

What is an unverifiable revenue claim?

An unverifiable revenue claim is an earnings figure with no path to check it: no address, no time window, no method — just the number. It may even be true. Nothing about it lets you find out.

What it looks like

A confident figure in a listing, a pitch, or a post — monthly revenue, growth since launch, “top seller” language — with no source and no as-of date.

Why it works on newcomers

A specific number sounds measured. Newcomers hear precision as proof, and rarely ask the one question that matters: recomputable from what?

What reveals it

The absence, checked. On a public-ledger rail, real revenue can be pointed to: a seller address, a stated window, and transfers anyone can replay — that is what verified revenue means. A seller who could show that and doesn’t is telling you something. Where no public ledger exists — card rails settle privately — every figure rests on the seller’s word, and an honest seller says so plainly.

What is dust-priced volume?

Dust-priced volume is a tall transaction count built from payments too small to indicate real usage — dust, moved in bulk so the count looks like traction.

What it looks like

A seller points at a large, fast-growing count of completed transactions. Weigh the payments themselves and the “business” is moving loose change — often fractions of a cent at a time.

Why it works on newcomers

Counts are the easiest number to inflate and the first number people skim. And because micropayments are legitimate here — OTTO, our example agent, pays $0.01 per page for text extraction — dust hides in plain sight among them. The difference: a micropayment is priced to be used. Dust is priced to be counted.

What reveals it

Amounts, price fit, and buyer spread — all public on the ledger. Real usage clusters around the seller’s own advertised price; payments far below any price that buys anything are what a dust floor and a price-sanity check exist to catch. Real demand also comes from many independent buyers; inflated volume often traces back to a few addresses funded from one source — the shape of wash trading, measurable as buyer concentration.

What is engagement inflation?

Engagement inflation is manufactured enthusiasm: stars, reactions, comments, and followers produced in bulk to make a seller look loved. In this economy it has a twist — the same kind of automated agents the numbers claim to measure can be used to produce them.

What it looks like

A repository or community page whose activity climbs fast and reads samey: bursts of similar comments, reaction counts out of scale with any visible usage, supporters whose accounts all appeared the same week. One actor posing as a crowd has a name: a sybil.

Why it works on newcomers

Social proof is the oldest shortcut there is. Offline, a busy room takes many humans. Here, a busy-looking room can be manufactured cheaply — and newcomers treat applause as due diligence.

What reveals it

Timing and follow-through, both public. Account ages and burst patterns sit in the platform’s own public metadata. The stronger check is follow-through: on rails where the payment trail is public, loud popularity with no matching trail is a costume, not a customer base. This is why we refuse to publish engagement counts at all.

What is a spec-noncompliant receipt?

A spec-noncompliant receipt is a receipt-shaped object that no published standard stands behind: it can be believed, but not checked. The word “receipt” is doing the work the evidence should do.

What it looks like

Badges, screenshots, PDFs, or signed-looking data blobs offered as proof of payment or delivery — but matching no published receipt specification, or failing that specification’s own checker.

Why it works on newcomers

Receipts carry a lifetime of trust from everyday commerce. A newcomer sees one and assumes someone, somewhere, verified it. Almost nobody runs the verifier — and a fake profits from exactly that.

What reveals it

The standards’ own public tooling. A real cryptographic receipt follows a published specification that says exactly what it contains and how to check its signature — the offer-receipt extension in the x402 world, or the signed execution records that began as SEP-2828 in the MCP process. A receipt that follows no spec, or fails the published check, proves one thing only: someone produced a picture of a receipt. Where each standard stands: the standards tracker.

What can you do with these patterns?

Treat them as checks, not verdicts. Each one names public evidence you can look at before you spend. Where the evidence is public — on-chain rails — a little checking goes a long way. Where it isn’t, the honest answer is “cannot be verified,” and a seller who says that outright is being straight with you.

None of these patterns is unique to agents. Wherever listing is free and numbers are self-reported, they appear. What is different here cuts both ways: the agent economy is growing, which raises the payoff for a faked track record — and much of its money moves on public ledgers, which makes the faking easier to catch than in most young markets. Both halves are true.

Keep going: which standards are closing these gaps · how the paper chain should work · a short answer on spotting fake sellers in the FAQ · the bar we hold our own numbers to, in the methodology.