The next release needs a review. A developer wants access to the deployment wallet. Someone has connected monitoring to Slack, but the person who set it up is away when an alert arrives.
Each question lands with the CTO, who also has a product to ship.
In my work with trading firms, funds and DeFi teams, I care about what happens after a security concern is raised. A finding can sit open for weeks if nobody knows who should decide, make the change or check the result.
Take a DeFi team with eight engineers, a live protocol and another release coming up. The CTO currently handles security alongside engineering.
The CTO has become the default security lead
Our team has an audit report with a few fixes still to verify. A new integration is being developed. The deployment process depends on one engineer, and the founder wants to understand who can move treasury funds.
I’d start by asking the CTO to show me the security work from the last few weeks. How much is waiting? Does the same question keep coming back? Can the team explain why the release is ready, or is everyone relying on someone else having checked it?
If the CTO can manage that work and only needs help with one integration, I’d scope a focused review. If each release creates another queue of unanswered questions, ongoing security leadership becomes a more useful option.
What a CISO would own
A CISO, or Chief Information Security Officer, leads the company’s security work. The role can be full-time or part-time.
You’ll also see the term virtual CISO, or vCISO. Providers use these names differently, so I’d read the responsibilities and availability in the proposal before comparing titles.
For our DeFi team, I’d want a named person to maintain the security priorities, coordinate reviews and bring unresolved decisions to the people who can make them. That person needs regular access to the CTO and enough context to understand what the team is changing.
Suppose the new integration relies on a price feed that the original audit didn’t cover. That question needs to reach the release decision before the team ships.
The security lead would arrange the technical review and follow up on its findings. Engineers would implement and test any required changes. The CTO would commit engineering time, and the authorized decision-maker would decide whether the release could proceed.
NIST’s Cybersecurity Framework includes roles and authority. Here, I’d name who can delay the release while the oracle review is open.
The work reaches beyond the contracts
I’d check who can merge a release or modify the deployment workflow. Then I’d follow the signing requests and the services that can change what users see in the frontend.
An attacker who gains access to one of those systems may have a route to funds without finding a new bug in Solidity.
For the upcoming release, I’d identify who can change the production build and which permissions the deployment wallet retains. I’d also check whether treasury and monitoring responsibilities have named owners.
The signer-security article follows one of these paths in more detail: a securely stored key can still receive a dangerous request from a compromised bot.
An audit firm can review its agreed code and report findings. The team still needs to track changes outside that scope, arrange the next review and verify that fixes reached the deployment. That’s recurring work the security lead could coordinate.
Agree who does what before starting
For this team, I’d write down the working arrangement before adding another person to its Slack channels.
| Work | Security lead | Team responsibility |
|---|---|---|
| Choose security priorities | Explain the risks and maintain the proposed work order | Leadership agrees priorities and allocates time |
| Prepare a release review | Identify what changed, arrange reviews and track open questions | Engineers supply the code and tests. The CTO owns delivery |
| Fix a finding | Clarify the issue and arrange verification | Assigned engineers implement and test the fix |
| Change a wallet permission | Review the intended access and how to test it | Authorized account holders approve and execute the change |
| Respond to an alert | Define escalation and participate within agreed coverage | Named responders and signers remain available for the required actions |
I’d agree who handles customer security questions and compliance work, including any specialists the company needs.
If the engagement covers only occasional technical reviews, the proposal should name the security responsibilities that remain with the team.
What I’d do in the first month
For our team, the first month would start with the upcoming release. I’d ask for the architecture, deployment process, audit findings and access list, then walk through the systems with the people operating them.
If a former contractor still has deployment access, I’d assign its removal and verification. If the release changes an oracle dependency, I’d arrange that review before the team commits to a launch date.
The team would keep a short shared work list. Each item needs an owner, a due date and a way to tell whether the problem has been addressed.
By the end of that first month, I’d want the founders to be able to see which release checks are complete, which questions are still open and what help the team needs next.
Choosing products the team can operate
Suppose our eight-engineer team has budget for monitoring and employee-device protection. I’d compare the work each option leaves with the team before asking the founders to approve it.
| Tool or system | Where it helps | What I would verify |
|---|---|---|
| Security monitoring with custom agents | Name the alert owner and test the supported dependencies and permitted response. | |
| An alternative onchain threat-monitoring workflow | Run the same incidents through a pilot and confirm purchased features and escalation. | |
| Employee-device investigation and response | Check the team’s operating systems, licence and available trained responders. | |
| An alternative endpoint investigation and containment workflow | Compare fleet coverage and the safe isolation drill on the same devices. |
I’d compare the two monitoring candidates against the protocol’s scenarios and the two endpoint candidates against employee devices. Existing infrastructure and the people available to run it would influence the choice. I haven’t run those product pilots for this article.
The CISO should assign configuration, health checks, rule changes and backup response. If nobody can take those duties, I’d include operating help in the proposal or reduce the scope to something the team can sustain.
What should be different a few months later?
I’d use the following releases to make the checks repeatable. The team should know when a change needs another review, who verifies deployment permissions and where unresolved issues go before launch.
A response exercise would help test the arrangements. For example, suppose monitoring reports an unexpected treasury transfer. Can the responder identify the transaction, reach an authorized signer and determine which actions are available?
A missing contact or wallet restriction that blocks the response goes onto the work list.
The DeFi monitoring guide explains how to follow an alert through to a verified response.
I’d measure progress through completed changes, such as removed deployment access or a dependency reviewed before release. I’d also track overdue findings and whether the same failure returns after a fix.
Part-time needs a clear calendar
Before signing, I’d agree how much time is reserved, which meetings need attendance and how requests are prioritized. The team also needs to know what happens when the planned work exceeds that time.
The scope should name any hands-on implementation, independent audit work or specialist support included in the fee. It should also explain how additional work is approved.
When I’d recommend another arrangement
A specific codebase or integration with a clear review deadline may suit a focused engagement. An existing security leader may only need another specialist for wallet permissions or a protocol change.
At the other end, a company may need someone involved throughout the working week, managing a security team and making frequent operational decisions. I’d consider a full-time hire when the recurring workload and required availability exceed what a part-time engagement can support.
For our eight-engineer team, I’d begin with the actual queue and the next release. If the founders can reserve engineering time, give the security lead access to decisions and agree workable coverage, a part-time arrangement is worth considering.
If the engineers have no capacity to fix anything, I’d address that constraint before adding more reviews.
You can use the security responsibilities worksheet to put names against the work and identify what you need from an engagement. If these decisions keep returning in your team, get in touch about ongoing Web3 security support.