Security at QoreChain
QoreChain was built for adversaries that do not exist yet. The same discipline applies to the ones that do: post-quantum cryptography at every layer, a standing bug bounty, responsible disclosure, and independent review.
Quantum-safe by construction
Full-stack post-quantum cryptography
ML-DSA-87 (Dilithium-5, FIPS 204) for signatures, ML-KEM-1024 (Kyber, FIPS 203) for key encapsulation, SHAKE-256 for hashing. Applied to consensus, transactions, bridges, and state verification, not bolted on.
NIST Category 5
QoreChain targets the highest NIST security category, equivalent to AES-256, for 128-bit post-quantum security across the protocol stack. The standards it implements were finalized by NIST in August 2024.
Wallet security by default
QoreX, the official wallet, is non-custodial: keys are generated and stored only on your device, every QOR transfer is signed with mandatory ML-DSA-87, and the apps collect no analytics and no personal data.
A 67,500,000 QOR bug bounty
Opened at mainnet launch on 7 June 2026 and active for the lifetime of the network. On 1 September 2026 it opens as a structured public programme, Break QoreChain, with a defined season, a published scope, fixed reward tiers and a settlement date.
The pool, and how to check it yourself
The bug bounty allocation is held in a dedicated on-chain escrow. You do not have to take our word for the balance.
qor1c60gmsq68jpx8yzjj0pagl7kkggs4zc8mrfkkmcurl -s https://api.qore.host/cosmos/bank/v1beta1/balances/qor1c60gmsq68jpx8yzjj0pagl7kkggs4zc8mrfkkmThe escrow holds 67,500,000 QOR and releases over 48 months under the published tokenomics. The first season of Break QoreChain, from 1 September 2026 to 28 February 2027, runs on the tranche released at listing: 10,125,000 QOR, distributed in full at settlement.
The pool is denominated in QOR and held on-chain. Every payout is made in stablecoin, converted from the pool at the time of payment. You are paid in something you can spend.
We convert QOR to stablecoin only to pay a validated report, from the address above, and we publish the transaction hash with every payout.
This pool is funded separately from the insurance fund. The insurance fund exists to make users whole and is not spent on anything else.
Rewards
| Severity | Total | Triage decision within | How it is paid |
|---|---|---|---|
| Critical | 5,000 USDT | 21 days | 1,500 within 30 days of the triage decision, 3,500 on mainnet deployment of the fix or 180 days after the triage decision |
| High | 2,000 USDT | 21 days | 600 within 30 days of the triage decision, 1,400 on mainnet deployment of the fix or 180 days after the triage decision |
| Medium | 750 USDT | 60 days | In full, at the season settlement |
| Low | 250 USDT | 60 days | In full, at the season settlement |
| Insight | 100 USDT | 60 days | In full, at the season settlement |
These windows apply to reports received from 11 November 2026, 00:00 UTC. A report that reached us earlier keeps the terms published when it arrived: first payment within 14 days for Critical and High and 21 days for Medium, Low and Insight, the remainder on mainnet deployment of the fix or 90 days after the triage decision.
The triage clock starts the moment your report reaches our inbox, not on a date we choose, unless the report is queued under the ten report limit described below. Critical and High are triaged and paid first because they are the ones that can cost users money.
Our triage is done by a named human being rather than a queue, and we would rather tell you what that means than pretend otherwise. If more than twenty five valid reports reach us in any rolling thirty day period, we will say so on this page, and the Medium, Low and Insight triage window extends to 90 days for reports received after that notice. Critical and High are never extended, and no extension is ever applied to a report already in our hands.
For Critical and High, if the fix is not deployed to mainnet within 180 days of the triage decision, the remaining balance is paid in full regardless. Your reward does not depend on our engineering schedule. A Medium, Low or Insight report triaged after the settlement date is paid within 30 days of its triage decision.
Every valid report also earns a share of the season pool, distributed at settlement on 28 February 2027 in the proportions below. The pool is distributed in full whether or not anyone finds a critical. Categories with no findings are redistributed proportionally.
| Category | Share of the 10,125,000 QOR season pool |
|---|---|
| Critical | 40% |
| High | 25% |
| Medium | 15% |
| Low | 5% |
| Insight and hardening | 10% |
| Participation | 5% |
The participation share is split between everyone who submitted at least one valid original report. A duplicate of a finding already reported does not count toward it.
No single researcher receives more than 10% of the season pool, 1,012,500 QOR, across all categories. Anything above that limit is shared among the other eligible reports in the same proportions, so the pool is still distributed in full. These pool rules apply to the settlement of the current season.
Severity classification
- Critical. Direct theft or permanent freezing of user funds, unauthorised minting or supply inflation, consensus halt or chain split reachable by an unprivileged actor, forgery or bypass of the post-quantum signature path.
- High. Theft or freezing of funds under preconditions the attacker does not fully control, privilege escalation to validator or module authority, permanent denial of service on a core module, key recovery or predictable key generation in a production signing path.
- Medium. Temporary denial of service, incorrect state transitions without direct financial loss, authorisation flaws limited to non-financial data, cryptographic weaknesses without a demonstrated exploitation path.
- Low. Issues requiring unusual configuration or privileged access, information disclosure of non-sensitive data, defects with clear mitigations already available to operators.
- Insight. Hardening opportunities, defence-in-depth improvements, and documented deviations between shipped behaviour and published specification.
PQC implementation issues, including side channels, carry the highest priority within their severity band.
What a Critical requires
A report is classified Critical only when all of the following are true:
- The proof of concept runs against the current official mainnet release, on a local devnet or the public testnet, with no modified code, no privileged keys, no cooperation from validators and no more than one third of the stake.
- The impact is shown, not described: the proof of concept actually moves, freezes or creates funds, or halts or splits the chain.
- We can reproduce it from the report alone, in a clean environment.
- The issue is not already known to us, not public, and not in a component announced as under remediation.
- Nothing was exploited on mainnet and nothing was disclosed before the disclosure window below.
A report that misses one of these is not rejected. It is classified at the highest band it does meet.
One root cause, one reward
Every variant of the same underlying flaw is treated as one finding, whichever component or entry point it is reached through. The reward goes to the first report that reaches us, as stamped on arrival; later reports of the same root cause are duplicates. Each researcher can have at most ten reports awaiting a triage decision at a time; further reports are queued in order of arrival and their triage clock starts when they leave the queue.
Scope
In scope from 1 September
- Consensus and block production, including staking, slashing and distribution
- The bank module, locked-coin accounting and escrow accounts
- The ML-DSA-87 signing path and post-quantum key registration, Go implementation
- The EVM execution lane, chain 9801
- Public RPC and API infrastructure
- The QoreX wallet
Added during the season
Components under active remediation are not in scope at opening. We do not open a bounty over code we already know is being changed. Each one is added, with an announcement, as its remediation lands: the CosmWasm lane, the bridge module, the dashboard application, the C and Python bindings in qorechain-pqc, and qorechain-lightnode.
Added at the end of the season
In the final two weeks of February, after the external audit, the SVM execution lane enters scope. It is the lane that broke us on 21 August, and it is the last one we open, only after everyone has had a chance to try it.
Out of scope
Volumetric and denial of service testing against production, social engineering of the team or the community, third party dependencies without a demonstrated impact on QoreChain, any testing on mainnet that affects real users or real funds, and automated mass submissions.
Proof of concept required
Every submission, at every severity, must include a working proof of concept against a local devnet or the public testnet. No proof of concept, no reward. This is not a formality; it is what lets us pay you promptly instead of arguing for a month.
Reporting and disclosure
How a report travels
- 1Report privatelyWrite to security@qore.network with a description, reproduction steps, affected commit or version, and your assessment of impact. Do not open a public GitHub issue for security reports.
- 2Acknowledgement within 48 hoursWe confirm receipt within 48 hours with a reference number.
- 3Triage decisionA person decides validity and severity within 21 days for Critical and High, and within 60 days for Medium, Low and Insight.
- 4PaymentFor Critical and High, 30% of the reward is paid within 30 days of the triage decision and the remainder on mainnet deployment of the fix, or 180 days after the triage decision, whichever comes first. Medium, Low and Insight are paid in one payment at the season settlement.
- 5DisclosureCritical vulnerabilities are targeted for a fix within 90 days. A summary of each valid report is published 30 days after its fix is deployed to mainnet, with credit to the reporter unless they prefer anonymity.
Safe harbour
Research conducted in good faith and within the scope above will not be met with legal action from the QoreChain Association, and we will not request that a third party bring one. Good faith means you did not access, modify or exfiltrate user data or funds beyond the minimum needed to demonstrate the issue, you did not test against mainnet in a way that affects real users or real funds, you reported to us privately, and you gave us the disclosure window above. If you follow those conditions and we later disagree about severity or eligibility, that is a disagreement about a payment, never a legal matter.
Separately, the Association has applied to adopt the Whitehat Safe Harbor agreement published by the Security Alliance (SEAL). That framework covers a different situation: it pre-authorises a whitehat to intervene during an active exploit in order to rescue funds. It is not a substitute for the paragraph above and it does not cover routine research. We will publish the on-chain adoption once it is registered.
Reports received before 1 September
Reports received before the programme opened are covered at full programme rates, regardless of the staged scope above.
Independent review
Third-party audit
An independent third-party security audit of the protocol is ongoing, with results to be published in the project data room when complete. The core protocol builds on the Cosmos SDK, and QoreChain's public interfaces, types, and protocol definitions are open source on GitHub.