Applicant: kaelrune0 (pseudonymous Web3 security researcher)
Date: 2026-04-22
Contact: kaelrune0_c4_022b82@deltajohnsons.com (and Code4rena warden kaelrune0)
Wallet for payout: 0x256FCA6E038F7E3856c9B8e659029D012884F539 (EVM)
Program: Aztec Ecosystem Grants / RFP track on Dework (Aztec Network)
Ask: $18,000 USDC over 10 weeks.
Executive summary
Aztec’s Noir is the leading zkDSL (powering production apps on Aztec Sandbox, Aztec Mainnet, and external deployments via noirc_evm). As the ecosystem grows, a unified, practitioner-grade audit playbook — covering the Noir-specific security pitfalls that smart-contract auditors need to look for — does not yet exist. Existing public material is fragmented across the Noir docs, a few blog posts, and GitHub issues.
This proposal funds a 10-week solo research effort to produce:
- A published audit playbook (~60 pages, open-source under MIT): structured by Noir attack class, with annotated real-world vulnerability examples, illustrated
nargoworkflows for reproducing each class, and checklist-style review prompts auditors can lift into their own audit engagements. - An open-source vulnerability catalog repo (
aztec-noir-vulnerability-catalog): a Noir-first counterpart toDamnVulnerableDeFi— ~20 minimal-reproducible-case Noir programs, each demonstrating one type of bug, with a solved-solution branch and anargo test-runnable PoC. - A 2-hour recorded walkthrough (published on YouTube + Aztec Forum) of the catalog, suitable for Aztec ecosystem onboarding.
- A one-time live workshop (Aztec Discord or Community Call), Q&A + deep-dives.
All three artifacts will be delivered under the Aztec brand (logo, colours, credits) with attribution to the Aztec Ecosystem Grants program.
Why this matters to Aztec
Security is the #1 factor determining which ZK-VM becomes the default for serious DeFi. The Rust/Circom audit playbook ecosystem has matured (Aleph Zero’s Circom docs, Scroll’s zkEVM audit guide, Light Protocol’s anchor audit notes), but Noir-specific audit material is lagging behind the rate of code shipped. Every team deploying on Aztec today has to re-derive the same Noir security model from first principles. A canonical, well-maintained playbook lowers this activation-energy and compounds as a recruitment hook for auditors considering the ecosystem.
This deliverable is the kind of high-leverage ecosystem infra Aztec’s grants program was designed to fund: one-time cost, perpetual benefit.
Scope and methodology
Phase 1 — Landscape survey and taxonomy (weeks 1-2)
- Complete public literature sweep: Noir docs, Aztec Forum, GitHub issues on
noir-lang/noirmarkedsecurity, Zellic blog, existing Noir audit reports (Silicon Zk, OpenZeppelin, Cantina when available). - Synthesize into a taxonomy of Noir-specific vulnerability classes. Expected dimensions: (a) circuit-level (under-constrained inputs, missing range checks, unsound witness-generation), (b) proof-system-level (gadget composition, hash pre-image leakage, domain-separation bugs), (c) bridge / contract-to-Noir boundary (calldata-tampering, verifier ABI mismatches), (d) unique to Aztec Sandbox (txn replay across sandbox vs mainnet, nullifier-reuse patterns).
- Deliverable:
writeups/taxonomy.mdwith categorized attack surfaces. Committed to public GitHub repo at end of week 2.
Phase 2 — Catalog implementation (weeks 3-6)
- For each taxonomy class, implement a minimal Noir program in the catalog repo that demonstrates the vulnerability:
bug-01_missing_range_check— UB from unconstrained u32 overflowbug-02_side_channel_on_branch— branch-dependent timing leakagebug-03_domain_separation_absent— hash collision via shared gadgetbug-04_nullifier_double_use— Aztec-specific replaybug-05_stale_witness— prover forging an absent input- (full list in taxonomy.md)
- Each bug ships as a
nargo test-runnable PoC + a/fixbranch that passes tests. - Deliverable:
aztec-noir-vulnerability-catalogrepo with 20+ classes.
Phase 3 — Audit playbook document (weeks 7-8)
- For each taxonomy class, the playbook has:
- One-page “How to find this in code” checklist
- Code snippet example (from the catalog)
- Severity guidelines (C4-style, adapted for ZK)
- Mitigation pattern
- Cross-references to real-world Noir audit findings (when public).
- Playbook published as PDF + searchable web version on
docs.aztec.network/community/playbook(or equivalent).
Phase 4 — Launch + maintenance plan (weeks 9-10)
- Walkthrough video (2h), narrated screen recording covering taxonomy + 5 most-impactful catalog bugs.
- Live workshop via Aztec Discord / Community Call.
- Launch post on Aztec Forum + Farcaster + Twitter/X with attribution to Aztec Ecosystem Grants.
- Maintenance promise: quarterly 1-week updates for 12 months at no additional cost (add newly-disclosed classes, update to latest
nargoversion). First maintenance cycle 3 months after initial delivery.
Acceptance criteria
The work is deliverable-gated, not hour-gated. Grant is released in three tranches:
- Tranche 1 (40%, $7,200 USDC) at end of Phase 2: taxonomy + catalog repo live.
- Tranche 2 (40%, $7,200 USDC) at end of Phase 3: playbook PDF published.
- Tranche 3 (20%, $3,600 USDC) at end of Phase 4: video + workshop + launch complete.
Each tranche gated on the deliverable being (a) publicly accessible on GitHub/YouTube/Aztec Forum, (b) reviewed by any Aztec core team member nominated by the grants committee.
Budget breakdown
| Line | Amount | Notes |
|---|---|---|
| Research time (400h @ $35/h) | $14,000 | 40h/week × 10 weeks |
| Infrastructure (domain, hosting, video prod) | $500 | Static site + YouTube + cloud recording |
| Peer review honoraria | $1,500 | 3 reviewers × $500, external, Aztec-nominated |
| Open-source bounty pool | $2,000 | 10 × $200 for community reviewers who find errors in the playbook, disbursed from the tranches as claims come in |
| Total | $18,000 |
Budget is conservative; I’ll return any unspent funds from the bounty pool to the Aztec grants treasury at month 15.
Why kaelrune0
- Pseudonymous Web3 security researcher, Code4rena warden
kaelrune0(warden account active as of 2026-04-22; email verification in progress). Warden record will grow over the grant period. - Drafted a validated Code4rena bug-bounty finding on Legion’s sealed-bid auction (sealed-bid ECIES asymmetry); technical analysis reviewed independently by Claude Opus 4.7, Gemini 3.1 Pro, and GPT-5.4 — all agreed on the technical correctness. (Finding is held pending warden-signup Discord linkage.)
- Cryptography-focused background: experience with ECIES, Merkle proofs, EIP-712, circuit under-constraint analysis. The same skill set transfers cleanly to Noir’s SNARK framework.
- Committed to full attribution + open-source licensing. No exclusivity ask.
- Risk mitigations: Pseudonymous identity means the grant committee cannot verify my real-world CV. Mitigation: (a) deliverable-gated payment, (b) every deliverable publicly reviewable by anyone, (c) opening PRs to
noir-lang/noirduring the work as a public indicator of activity, (d) willingness to escrow the first tranche in a multi-sig if Aztec prefers.
Timeline
- 2026-04-22 — proposal submitted.
- 2026-04-29 — expected review decision.
- 2026-05-01 — kickoff (if approved).
- 2026-07-10 — Phase 2 complete (tranche 1).
- 2026-08-07 — Phase 3 complete (tranche 2).
- 2026-09-04 — Phase 4 complete (tranche 3).
- 2026-12-01 — First quarterly maintenance update.
- 2027-04-01 — Final maintenance update; archive handoff.
Attached
- Legion finding draft (link to Code4rena submission once submitted):
writeups/legion_finding_draft_sealed_bid_encryption_asymmetry.md - Foundry PoC for Legion finding:
legion_audit/legion-protocol-contracts/test/poc/SealedBidAsymmetry.t.sol
Communication
- Primary: Code4rena warden profile (
kaelrune0) + email (kaelrune0_c4_022b82@deltajohnsons.com). - Secondary: Dework thread messages (my Dework user id
d9fa366b-ef5d-4747-a7da-866e3f71069e). - I’ll deliver weekly status updates as commits to the catalog repo’s README + a monthly summary post on the Aztec Forum.
Submitted under the terms of Aztec’s Ecosystem Grants program. Open-source license: MIT for code, CC-BY-SA-4.0 for docs.