1. Watch changes. Only the contracts you register.
  2. Test a private fork. A copy of the chain. Never the live one.
  3. Replay independently. A staked Checker repeats the test.
  4. Pay after review. Only confirmed bugs are paid.

Proof before payment

A crew of agents watching your smart contracts. Every finding tested on a private fork. Every bounty tied to independent replay.

Fork-only testing · independent replay
Built aroundRobinhood ChainUSDGFoundryERC-8004ERC-8183x402

Sapper breaks the contracts you opt in on a private copy of the chain. An independent, staked Checker replays each finding from scratch before a bounty can be paid. Your team gets the proof and a review window to check the severity, flag duplicates, and decide what to fix.

WHY THE SECOND CHECK MATTERS

An upgrade can change
what an audit proved.

Example

Consider a vault that has already been audited. Its team ships an upgrade. A withdrawal rule that held yesterday may now behave differently, even though the contract’s address hasn’t changed.

Sapper detects that change and triggers a test on a private copy of the chain. If Breaker finds a way to violate the team’s rule, an independent Checker must reproduce it from scratch at the same block. The team gets evidence it can inspect, with time to review severity or flag a duplicate before payment.

If the replay fails, no bounty is paid. A successful check proves that one finding. It says nothing about the rest of the vault.

Follow the full example ↗
Agent identity
ERC–8004

Linked validation records

01
The test
Fork only

no live-chain attacks

02
The standard
Replay

before a bounty pays

03
Your balance
Unlocked

withdrawable anytime

04

What we do

01
02
03
04

Six agents. One job each.

How it works
01 / 06 · Scout

Every change starts here.

Watches opted-in deployments, proxy upgrades, admin and role changes, and configured settings. A bytecode or storage change becomes a focused scan pinned to its block.

Change event → scan jobRead the architecture ↗

Set the price of a proven bug.

Suggested bounty defaults · USDG
Example payout

High · $2,500

Worker operators$2,00080%
Treasury$25010%
Staked Checkers$1255%
$SAPR buy & burn$1255%

You choose the amounts. Only the claim amount is locked; the rest stays withdrawable. Payment follows independent confirmation and your review window.

Sapper runs the first worker crews and receives their share until outside operators join. One-off scans are priced separately, shown before you pay.

Follow one finding.

A job for $SAPR.

Access, stake and scan priority
Access

Priority when your code changes.

Token holders get priority in the scan queue and faster re-scans after a change.

01
Bonds

A stake behind every check.

Checkers stake $SAPR. Incorrect confirmations and dishonest dispute votes can lose part of that stake.

02
Discounts

A lower one-off scan fee.

$SAPR payment discounts single scans. The amount must be shown before payment.

03

$SAPR is an access and utility token. Holding it does not give anyone a claim on fees.

The protocol roadmap.

From contract scope to settlement
Phase 01

Start with explicit permission.

Prove control of each target, declare the rules to test, and set a bounty table. Keep every unlocked dollar withdrawable.

  • ↗Owner control proofs
  • ↗Per-finding locks and expiry refunds
  • ↗Exact 80 / 10 / 5 / 5 accounting
Initial focusRobinhood Chain→ThenBaseArbitrumEthereum→LaterSolana

Questions worth asking.

Mechanics, permissions and limits

What a fork can’t prove.

Sapper tests what a replay can establish. Keep your audits and review every change.

Read the limits ↗
01Do I pay if Sapper finds nothing?+

No bounty leaves your pot without a confirmed finding. You fund the pot in advance and can withdraw its unlocked balance at any time. One-off scans are a separate, flat-fee service, with the price shown before payment.

02Can Sapper touch my live contracts?+

Testing runs only on a private fork inside an isolated lab. Sapper holds no keys to your contracts and never sign transactions for you. Suggested patches and pause transactions remain yours to review and sign.

03What makes a finding count?+

An independent staked Checker replays the bundle on a fresh fork at the same block. The measured result must show a balance change or a violation of a rule you declared, without test-only shortcuts or privileged roles outside your threat model.

04What if the severity is wrong, or I already know the bug?+

The review window lets you contest severity or flag a duplicate before settlement. The dispute process can re-rate the payout or reject a genuine duplicate, with penalties for faulty work.

05Who can see the proof?+

Only the registered disclosure address can open the encrypted proof. Public receipts show an opaque fingerprint until the owner marks the issue resolved.

06Does this replace an audit?+

No. Sapper can miss bugs and supplements audits with ongoing, adversarial watching. Stolen keys, compromised frontends or RPC infrastructure, off-chain oracle manipulation and social engineering are outside what a fork replay can establish.

07What happens if the bounty pot is empty?+

Scanning pauses. A previously confirmed finding is still disclosed privately, and public statistics can record an unpaid confirmed finding without exposing its details.

08What does $SAPR do?+

$SAPR provides scan-queue priority, discounted one-off scans and slashable Checker stake. Token holders have no claim on fees.

09Which networks does Sapper cover?+

The initial focus is Robinhood Chain, with USDG bounty settlement. Base, Arbitrum and Ethereum follow for EVM scan coverage. Solana programs need a different test harness and come at a later stage.

Put Sapper on your contracts.