The Bitget hack has produced two opposite answers to the same question.
When stolen funds reached THORChain, the protocol refused Bitgets request to block known attacker wallets and defended its permissionless design.
When similar funds reached NEAR Intents, NEAR intervened.
NEAR Intents said its SHIELD system identified more than $50 million in attempted transfers linked to the Bitget attacker. Most were rejected before execution. Roughly $166,000 passed through the system, while approximately $503,000 was frozen during execution and remains restricted pending legal and recovery procedures.
NEAR co-founder Illia Polosukhin defended the decision with a useful phrase:
permissionless does not mean neutral.
Intents Are Not the Same as Direct Blockchain Transactions
A user interacting directly with a base-layer blockchain broadcasts a transaction that validators process according to protocol rules.
An intent system works differently.
The user specifies an outcome, such as converting asset A into asset B.
Independent solvers compete or coordinate to fulfill that request.
That architecture can simplify cross-chain trading.
It also creates an execution layer between the user and the final settlement.
If solvers or routing infrastructure can refuse a request, the system is permissionless at one layer and selective at another.
This is why simple labels such as “censorship resistant” become less useful.
The correct question is:
which layer has the power to say no？
NEAR Intents Uses a Screening Layer
NEAR says its SHIELD system detected the Bitget-linked flows.
The reported results show several possible outcomes:
- transactions rejected before execution;
- a small amount processed;
- funds frozen during execution.
That means the system does more than passive monitoring.
It can affect execution.
This has clear security benefits.
A stolen asset can be stopped before it reaches a more difficult laundering route.
It also creates governance questions.
Who maintains the risk model？
Who labels an address？
Who can release frozen assets？
What happens when an address is incorrectly flagged？
Those questions matter as much as the technical ability to screen.
The THORChain Comparison Is Useful — but Not Perfect
The industry immediately compared NEAR Intents with THORChain.
THORChain refused targeted blocking.
NEAR Intents intervened.
But the systems are architecturally different.
THORChain uses shared liquidity and threshold-signature vaults.
NEAR Intents uses solver-based execution.
The control points are not identical.
Still, the comparison reveals something important:
“permissionless” does not describe one universal operating model.
Different protocols can be open to users while making very different choices around execution, sanctions, stolen funds and risk controls.
Selective Intervention Can Improve Recovery
From an exchange-security perspective, NEARs intervention is useful.
The Bitget attacker has been moving assets across native chains and swap services.
Every service that rejects or freezes identified flows increases friction.
Stablecoin issuers can blacklist their tokens.
Centralized exchanges can freeze accounts.
Intent systems can refuse solver execution.
Native blockchains may have no equivalent control.
The recovery network is therefore heterogeneous.
Incident responders need to know what intervention power exists at every hop.
The Same Power Can Create False-Positive Risk
The harder side is mistakes.
Address attribution is probabilistic.
Funds can pass through intermediaries.
A wallet can receive tainted funds without knowing the origin.
If a protocol freezes based on automated risk labels, legitimate users need a clear appeal and recovery process.
CoinDesk noted that NEAR Intents had not publicly explained who can authorize the release of restricted funds or how a wrongly flagged user would recover money.
That is not a minor UX issue.
It is part of the governance model.
If a system has the power to freeze, users need to understand the process around that power.
Why It Matters
Cross-chain infrastructure is becoming a spectrum rather than a binary choice between centralized and decentralized.
Possible models include fully neutral base-layer execution, shared-vault protocols with emergency controls, solver networks with transaction screening, issuer-controlled tokens with blacklists and centralized exchanges with account freezes.
Each model trades off:
neutrality, recovery, compliance and user protection.
NEAR Intents is important because it makes that tradeoff explicit.
Its position is not “we cannot intervene.”
It is closer to:
we can intervene, and we believe some intervention is justified.
That is a governance philosophy.
“Permissionless” Needs a More Precise Vocabulary
The industry needs better words.
A protocol can be permissionless to submit, permissioned to execute, permissionless to settle, centrally curated at the interface, decentralized at the validator layer and selective at the solver layer.
Those properties can coexist.
Reducing them to one word creates confusion.
For institutional adoption, that nuance may be a feature.
A bank may prefer permissionless settlement with controlled execution.
A privacy user may prefer the opposite.
There is no single decentralization template.
Risks and Counterarguments
The $50 million-plus figure refers to attempted flows identified by NEAR Intents, not funds successfully frozen.
Only about $503,000 was reportedly frozen mid-execution.
The exact attribution of all attempted flows depends on blockchain analytics.
Intervention can reduce laundering but cannot stop attackers from using alternative protocols.
And transaction controls can expand over time from clear theft cases to sanctions or other policy categories.
What to Watch Next
Watch NEAR Intents formal freeze policy, release and appeals procedures, additional Bitget-linked flows, solver concentration, false-positive incidents and regulatory attention.
The most important long-term metric is not how much was blocked in one hack.
It is whether the protocol can make intervention predictable, transparent and contestable.
FAQ
How much Bitget-linked activity did NEAR Intents identify？
It said more than $50 million in attempted transfers was linked to the hack.
How much was frozen？
About $503,000 was reportedly frozen during execution.
How much passed through？
About $166,000.
Is NEAR Intents fully censorship resistant？
The Bitget incident shows its execution and screening layers can reject or restrict transactions.
Why is this important？
It shows that “permissionless” systems can still have selective intervention points, and those points need explicit governance.