Add Sherlock as a Super Validator (max weight 4)
Please see below open for discussion:
CIP Number: [CIP-TBD]
CIP Title: Lifecycle Security for Canton NetworkAuthor: Zachary Pollock (Sherlock)
Type: Governance / Security
Status: Draft
Created: 2026-05-28
Sponsor: MPCH
Abstract
Add Sherlock as a Super Validator with a maximum weight of 4.
Sherlock’s Lifecycle Security solution will strengthen the security posture of the Canton Network by addressing critical gaps that point-in-time audits cannot cover. This benefits both the core Canton codebase (built on Daml and Scala) and the broader ecosystem of institutional deployments.
Sherlock commits to delivering Lifecycle Security for the Canton Network by:
-
Hosting and fully managing a triaged bug bounty program covering all Canton Network core code (Daml, Scala, and related components).
-
Creating and administering a Security Initiative Fund, funded by Sherlock's SV rewards, to sustainably scale security investment as the network grows.
-
Developing and maintaining Daml- and Scala-specific iterations of SherlockAI — a white-glove, human-in-the-loop AI-assisted security tool — delivering regular AI-assisted audit reviews across Canton core code and ecosystem deployments, with human expert triage and actionable fix recommendations on every scan.
-
Driving adoption of Daml security expertise across Web3 through sponsored developer events, workshops, and audit contests — expanding the number of Daml-certified and Daml-skilled security researchers actively engaged in the ecosystem.
These initiatives will create a self-reinforcing security flywheel that raises the bar for institutional-grade confidence on Canton.
All net SV rewards above Sherlock's operational costs will flow into a Security Initiative Fund for the benefit of the Canton Network.Lifecycle Security Visualization
Motivation
Point-in-time audits alone have repeatedly failed to prevent major incidents in DeFi and on-chain protocols. In 2025, Web3 suffered $3.4B in losses from hacks, demonstrating that security failures often arise not just from missed audit findings, but from the absence of continuous, lifecycle-spanning security coverage.
The Canton Network is purpose-built for regulated institutional finance and is already attracting participants such as Visa, Broadridge, Tradeweb, DTCC, Nasdaq, and Circle. For a network operating at this tier, Lifecycle Security is not optional — it is a prerequisite for long-term adoption by entities that require rock-solid, auditable, and sustainable security posture.
The gold standard today is demonstrated by Aave V4's launch, which followed the full Sherlock Lifecycle Security journey: early consultation, SherlockAI development-time scans, multiple rounds of human audits, a large-scale audit contest as the final gate before mainnet, and a Sherlock-hosted bug bounty for post-launch protection. Canton deserves the same standard.
Why Sherlock
Founded in 2021, Sherlock is Web3's leading full-lifecycle security firm, supporting protocols responsible for more than $250B in active TVL and having identified more than 1,500 critical vulnerabilities. Sherlock is uniquely suited to fill Canton's security gap:
-
Lifecycle coverage: The only firm with an integrated platform spanning development (SherlockAI), pre-launch (audits + contests), and post-launch (bug bounties + monitoring via existing SV, Hypernative).
-
Proven at institutional scale: Flagship clients include the Ethereum Foundation, Aave, Morpho, Sky, Centrifuge, Usual, M^0, Cosmos, and Babylon.
-
Ecosystem grant leadership: Over $2M in security audits subsidized on Optimism, $800K+ on Arbitrum; whitelisted provider for Uniswap, Solana, BNB, and others.
-
Largest security researcher community: 10,000+ security engineers have interacted with Sherlock's audit contests and bug bounty programs over the past five years.
-
Institutional relationships: Sherlock has established relationships with security and engineering teams at Stripe (Tempo), Ripple, Robinhood, Aave, Morpho, Centrifuge, Maple, and others.
-
Proven AI Scan Model - Grants awarded by Ethereum & Arbitrum: The human-in-the-loop AI audit scan solution proposed for Canton is not new, it mirrors the exact program Sherlock is actively delivering for Ethereum Foundation and Arbitrum Foundation. Both organizations have contracted Sherlock to provide AI-assisted scans with human expert triage for projects building on their respective chains. Canton's offering is modeled directly on this established program, meaning the tooling, methodology, and operational processes are already battle-tested at the ecosystem level rather than being built from scratch.
-
Existing Canton relationships: Sherlock's SV node will be operated in partnership with MPCH, an existing Canton Super Validator (CIP-0017), using MPCH's cold key management infrastructure and making Sherlock the first new Super Validator with true cold key storage capability. Sherlock also has an existing working relationship with Hypernative (CIP-0076), Canton's real-time monitoring Super Validator, further embedding Sherlock within the Canton security ecosystem ahead of this proposal.
No-Profit Commitment
Sherlock does not intend to profit from Super Validator rewards. All rewards above operational costs will be directed to Canton Network security initiatives. The table below summarizes the monthly cost structure, to support a stronger, more resilient security posture for the entire network:
MONTHLY
EXPENSES
Bug Bounty Hosting
$4,000
AI Maintenance & Management
(benchmarking/optimization/maintenance)$10,000
AI assisted security audits
(compute + labor cost + coordination)$10,000
Daml Engineer Adoption
(workshops, events, bounty competitions)$4,500
MPCH SV Fees
$12,000
Network SV Fees
$6,000
MPCH Cold KMS
$3,500
TOTAL = $50,000
Sherlock commits to converting a maximum of $50,000 of $CC to $USD each month to cover operating costs, then allocating the full remainder to the Security Initiative Fund. This structure will be maintained long-term, with a 5% annual increase to account for rising costs.
Deliverables for Full SV Reward
Deliverable
Acceptance Criteria
Deadline
Weight Earned
SV Node Live
SV node deployed, operational, and connected to the Canton Network
+30 days from CIP Approval
0 (prerequisite)
Network Security Initiative Fund
Security Initiative Fund structure approved in collaboration with Canton Foundation to enable next steps
+30 days from CIP Approval
+0.5 (Total: 0.5)
Canton Bug Bounty Program Launch
(1) Sherlock completes a human audit of the Splice codebase to establish auditor context and identify/resolve any existing pre-bounty vulnerabilities.
(2) Canton bug bounty program fully staged in Sherlock's client-side platform and ready to be pushed live, pending final approval.
+90 days from CIP Approval
+1.0 (Total: 1.5)
Bug Bounty Fully Operational
Canton bug bounty program publicly hosted and accessible for submissions on Sherlock's public-facing bug bounty page at sherlock.xyz.
+120 days from CIP Approval
+0.5 (Total: 2.0)
Daml / Scala-Specific SherlockAI Tooling
Daml and Scala-specific SherlockAI model released. Trained on Sherlock's Daml audit history, proprietary models, and public audit data. Delivering a minimum of 30 human-in-the-loop AI audit reviews per calendar quarter - applied to a combination of Canton Network core code & codebases deployed on the network. Each review combines AI audit scan with human expert triage + human fix recommendations.
+150 days from CIP Approval
+0.5 (Total: 2.5)
Security Initiative Fund: First Distribution
First security initiative funds distributed to a Canton ecosystem project or via a live valid bug bounty program payout on Canton code.
Sherlock has a direct line to the Canton Network security committee, via our relationship with MPCH CTO, Richard Domikis, and will consult with the security committee to best understand the immediate low-hanging needs for this program, and will implement the first phase based on this feedback.+180 days from CIP Approval
+0.5 (Total: 3.0)
Daml Security Researcher Growth: Milestone 1
50 unique security researchers have identified a Medium or above vulnerability in Daml codebase, via Sherlock audit, audit contest, or bug bounty.
+2 years from CIP Approval
+0.5 (Total: 3.5)
Daml Security Researcher Growth: Milestone 2
100 unique security researchers have identified a Medium or above vulnerability in Daml codebase, via Sherlock audit, audit contest, or bug bounty.
+3.5 years from CIP Approval
+0.5 (Total: 4.0)
Note: The Researcher Growth milestones (Total 3.5 and 4.0) form part of the maximum weight of 4 committed under this CIP. Rewards earned from these two milestones will be retained by Sherlock — rather than directed to the Security Initiative Fund — to be reinvested directly into Daml developer adoption activities, security audit contests, and ecosystem engagement initiatives.
Deliverables & Security Flywheel Described
Sherlock will design, implement, and operate a complete security flywheel that directly secures Canton core infrastructure while delivering overflow benefits to the network ecosystem.
-
Bug Bounty
AI-era Bug Bounty Programs Require a Different Approach
-
Sherlock invented a novel "stake-to-submit" feature 6 months ago, where researchers are required to stake a slashable $250 in order to submit a bug. Nearly every Web3 bug bounty platform has now copied this model. It's a great way to filter AI slop, while still maintaining engagement. We still receive AI spam but the signal-to-noise ratio is strong.
-
Sherlock has an AI judging solution that has been live in production for the past 2+ years already, used in our audit contests and now being applied to high volume bug bounty program
-
Sherlock's bug bounty program includes a white glove triage service, meaning the engineer triaging your bugs is hand-selected because they have personally audited your codebase. This provides accurate triaging, and provides that individual with context to identify if bug submissions require deeper investigation (for example: if the actual submission is incorrect, but something similar could be a threat). If the bug is invalid, the $250 is paid to the triager - which makes the model economically viable.
-
Finally, SherlockAI is used as an offensive tool to hunt bugs in your own codebase so these bugs can be remediated before bug bounty hunters find them. A portion of our SV's included monthly SherlockAI compute credits are intended to be provided to DA for this purpose.
Bug Bounty Reward Structure
The following is a proposed reward structure, subject to review with the Security Committee. Rewards will begin conservatively until the Security Initiative Fund reaches sufficient balance to support higher payout tiers.
Severity
Reward Range
($USD denominated, paid in $CC or stablecoin)Critical
$25,000 – $500,000
High
$5,000 – $25,000
Medium
$5,000 (flat)
Bug Bounty Operations
DA will designate a single point of contact or committee responsible for collaboration with Sherlock's bug bounty operations. Sherlock will lead bug bounty value determination, structure, triaging, and administration. The DA representative will be responsible for bug fixes and for providing Sherlock with technical feedback on vulnerability validity, severity relative to intended design, and vulnerability impact. A Sherlock Daml auditor will perform a post-fix code review after each code update related to a bounty submission.
If the Security Initiative Fund balance is ever insufficient to cover a bounty payout, Sherlock will collaborate with the Security Committee and Treasury Committee to structure a fallback payment schedule.
-
SherlockAI: Human-in-the-Loop Model
Sherlock will develop and maintain a specialized version of SherlockAI tailored for Daml and Scala (the core languages of Canton).
Many companies have developed AI auditing tools, but the issue is that most of them identify issues without providing fixes or a follow up security review. The solution is human-in-the-loop: AI performs initial scans, followed by Daml-specialist auditors who triage findings, eliminate false positives, consult with developers, and provide actionable fix recommendations.Public Good (Base Offering):
Available to Canton core contributors and Security Initiative Fund recipients. Sherlock commits to delivering a minimum of 10 human-in-the-loop AI scans per month. AI performs initial scans; Daml-specialist auditors then triage findings, eliminate false positives, consult with developers, and provide actionable fix recommendations.Paid Offering:
Available to all Canton Network participants. Any codebase deployed on Canton Network receives a 30% discount.-
Security Initiative Fund
At the end of each quarter, any unused portion of SV rewards above operational costs will automatically transfer to a transparent Security Initiative Fund wallet overseen by Sherlock. These funds will support security initiatives such as audit contests for critical code, subsidized audits for strategic deployments, human-assisted AI audit scans for ecosystem projects, and other high-impact contributions. This ensures that Sherlock's SV rewards, above operational costs, are directed back into network security initiatives, reconciled quarterly, to be reinvested into chain security activities-
Daml Security Researcher Growth
Sherlock commits to hosting a minimum of one developer-focused Daml security event per calendar quarter. These events will include workshops, audit contests, and hands-on training designed to grow the number of Daml-certified and Daml-skilled security researchers across Web3.
Appendix
Appendix - Transparency & Oversight
All Security Initiative Fund flows, grant awards, and bounty payouts will be publicly reported and verifiable on-chain where possible. Sherlock will coordinate with the Canton Foundation, Tech & Ops Committee, Tokenomics Committee, and Security Committee as needed. This CIP is indefinite in duration, subject to standard Super Validator review and governance processes.
Appendix — SV Reward Mechanics
-
- An extraBeneficiary PartyID associated with the ‘escrowed’ Super Validator will be setup by the Foundation, or another SV node operator approved to provide SV rewards escrow services, with an SV Weight at the maximum earnable weight.
- The Applicant is responsible for coordinating the process of setting up the escrowed weights with the GSF and the operator of the SV node.
- The Applicant is responsible for all costs associated with the operation of the representative SV
- The representative SV will NOT mint rewards on a block by block basis
- All representative SV rewards will go to the Unclaimed Rewards pool
- Applicant is required to present proof of successful completed milestones to the Tokenomics Working Group
- Applicant is required to present a calculation for number of Canton Coin it should earn for meeting the requirements of the milestone
- If the Tokenomics Working Group agrees the milestone has been met and agrees with the calculation, an announcement will be sent via the Tokenomics-Announce mailing List
- ⅔ of Super Validator Operators will then assign a portion of the Unclaimed Rewards to be minted by the Applicant’s Validator
- ⅔ of the Super Validator Operators will update their configurations to allow Applicant to takeover a portion of their SV Weight on a go-forward basis
- If any milestones and associated rewards are not achieved by the deadline
- Applicant will be notified they have not met a deliverable by the Foundation
- Remaining SV Weight on the representative SV will be removed from the SV Operator configs
- The Tokenomics Working Group will make a recommendation to the SVs on what to do with the Unclaimed Rewards
- An extraBeneficiary PartyID associated with the ‘escrowed’ Super Validator will be setup by the Foundation, or another SV node operator approved to provide SV rewards escrow services, with an SV Weight at the maximum earnable weight.
-
- Hello,--
I would like to share Sherlock's CIP proposal. MPCH is our sponsor and we would like to post this for vote. MPCH has already shared this proposal with the Security Subcommittee and indicated strong interest and support.
Can you please advise on our best next steps towards formally sharing this with voting members?
Thank you so much!
Zack
