Rug Pull, Honeypot, Farming: How Solana Token Scams Actually Work

Solana produces new tokens faster than anyone can read about them. Thousands per hour, most of them worthless by the next morning.

What kills them is rarely indifference. It is design. Almost every way a token takes money off buyers falls into one of three shapes: the liquidity gets pulled, the sell button stops working, or the supply was quietly owned by the launcher before the chart even existed. Rug pull, honeypot, farming.

They fail differently and they leave different fingerprints. Knowing which one you are looking at is the difference between a scan you can act on and a risk score you just stare at.

A rug pull removes a token's liquidity so it can no longer be sold at a real price. A honeypot lets you buy but blocks the sell. Farming means insiders held most of the supply from the first block and sold into your buy. All three leave on-chain evidence before the price moves.

What a rug pull actually is

A rug pull is the exit. Whoever launched the token withdraws the liquidity pool, or dumps a supply they were sitting on, and the price goes to nothing in a single block.

The mechanics are simpler than the name suggests. A token trades against a pool of paired liquidity, usually SOL. That pool is the only thing giving the price a floor. If whoever controls the LP tokens can withdraw them, the floor is optional.

Two versions are worth telling apart. A hard rug is written into the launch: liquidity never burned or locked, mint authority left live, and the exit happens the moment there is enough on the other side to be worth taking. A soft rug is slower. The team keeps posting, keeps shipping just enough to look alive, and sells into every rally until nothing is left. Hard rugs are straightforward fraud. Soft rugs sit in a grayer area, which is exactly why they run longer.

The scale is not a rounding error. Rug pulls account for roughly a third of all crypto scam losses, and Solana and BNB Chain carry most of the volume because deploying a token there costs almost nothing. On pump.fun, the graduation rate, meaning the share of tokens that gather enough liquidity to migrate to a DEX, had collapsed to about 0.26 percent by mid-June 2026. Around 997 of every 1,000 launches never get that far.

A rug pull is when a token's creators withdraw its liquidity or dump their holdings, leaving holders with no realistic sell price. Hard rugs happen in one transaction; soft rugs bleed value out gradually while the team keeps promoting the project.

Honeypots let you in and keep you there

A honeypot inverts the trick. The liquidity stays put, the chart looks healthy, buys go through normally. Selling is the part that fails.

On Solana the usual mechanism is freeze authority. When a token is created, the creator can keep the ability to freeze any holder's token account. Buys keep working, the price keeps climbing on real money, and when the creator is ready they freeze everyone else and sell their own bag. You watch the whole thing happen holding an asset you cannot move.

On EVM chains the same outcome gets written into the contract instead: a whitelist check on transfers, a sell tax that jumps to 99 percent after launch, a blacklist function that quietly grows. Different code, same room with no door.

This is the one that is genuinely checkable in seconds. Freeze authority is public. If it reads null, it has been revoked and nobody can lock your wallet. If there is an address sitting there, someone kept the key, and the only thing protecting you is their restraint.

A honeypot token allows buying but blocks selling, usually through a live freeze authority on Solana or a transfer restriction written into the contract on EVM chains. Confirming that freeze authority is revoked rules out the most common version.

Farming is the one most people miss

Farming is not an event. It is a business model.

The setup runs like this. Before a token is public, the deployer funds a set of wallets and has them buy in the same block as the launch, ahead of anyone who could physically have seen it. Bundlers push those buys through together, so the chart opens with what looks like organic demand from dozens of independent participants. It is one person.

The numbers are documented. One month of pump.fun activity showed more than 15,000 launches involving over 4,600 sniper wallets and 10,400 deployers running exactly this pattern. 87 percent of those snipes were profitable. More than half were fully exited in under a minute, and in over 90 percent of cases the wallet sold everything in one or two swaps. None of that is investing. It is a harvest, and the crop is whoever bought thirty seconds later.

The second form of farming is repetition. One deployer runs the same playbook across dozens of tokens, swapping the name and the art each time. Each launch pulls in a modest amount and dies, and the operation is profitable on volume rather than on any single hit. This is why a creator's history tells you more than a token's chart does. A wallet that has launched forty tokens and rugged thirty-eight of them has already told you how number forty-one ends.

Wash trading finishes the picture. Bot volume to hold a trending slot, bought engagement to make a Telegram look busy. None of it is demand. All of it is bait.

Farming is when insiders take most of a token's supply at launch through bundled or deployer-funded wallets, then sell it into real demand. Unlike a rug pull it needs no dramatic exit, because the profit was built in during the first block.

Why a clean contract check is not the answer

Here is the part that costs people real money. A token can pass every contract check and still be a farm.

Mint authority revoked. Freeze authority revoked. Liquidity burned. Three green ticks, and 40 percent of the supply still sitting across wallets that were funded from the same source three minutes before launch. Nothing in the code is wrong. The distribution is the scam.

Contract flags answer one question: can the code hurt you. They say nothing about whether the holders are the same person wearing different wallets. On Solana, where launching is cheap and the playbook is fully industrialized, that second question is usually the one that decides how your trade ends.

The checks that change a decision

A useful pre-buy read is short. These are the things that should actually change what you do:

• Mint authority: revoked, or can more supply be printed after you buy
• Freeze authority: revoked, or can your wallet be locked out of selling
• Liquidity: burned, locked with a real unlock date, or withdrawable right now
• Holder concentration: what the top wallets control once you exclude the LP
• Funding clusters: wallets that look unrelated but were funded from one source
• Sniper and bundle share: how much supply was bought in the launch block
• Creator history: how many tokens this deployer has launched, and how those ended

Any one of these on its own proves very little. A high top-holder number can be a legitimate treasury. A first-time deployer is not automatically a scammer. The signal lives in the combination, and in whether the wallets behave like separate people or like one operator with a script.

Running the check without doing it by hand

You can pull all of this yourself out of a block explorer if you have the time and know where to look. Most people are deciding in under a minute, which is what the tooling exists for.

xAxios AI's Rug Check runs 150+ wallet-level checks on any Solana token and rolls them into one score across eight signal categories, covering snipers, bundlers, insiders and bots alongside the standard contract flags. It also surfaces the creator's launch history, so a serial deployer shows up before you buy instead of after. The browser extension runs the same scan directly on DexScreener, DexTools, GeckoTerminal and OKX Web3, so the check happens on the page you were already looking at.

No scanner is a guarantee, and running a second one is still the right habit for anything you are putting real size into. Different tools catch different things, and the disagreements between them are often the most informative part.

Three patterns cover nearly all of it. Liquidity that can walk out, a sell that can be switched off, and supply that was never really distributed. Reading those three before you buy takes about a minute, which is a minute more than most of the money lost to them ever got.