Secure software delivery is no longer optional — it's what separates products users can trust from the ones that end up in post-mortems. In 2025, Hacken ran an SSDLC (Secure Software Development Life Cycle) maturity survey across our customers and partners to take an honest look at how blockchain teams are actually building software today.
The survey covered over 20 questions spanning five domains: governance, security by design, engineering controls, verification & testing, and operational resilience. Here's what the data revealed.
The foundations are solid. The gaps are dangerous.
Most teams showed strong habits where it's easiest to build them: 93% actively monitor production, 87% follow Agile principles, and over 80% have peer code reviews and MFA in place. That's genuinely encouraging.
But dig deeper, and a different picture emerges.
Around 30% of teams lack formal technical security training. General awareness training — phishing simulations, social engineering drills — doesn't equip developers to prevent smart contract exploits or catch code-level vulnerabilities. And no amount of external auditing compensates for a team that can't identify risks in its own code.
Then there's the audit gap. 71% of teams have run a security audit, but nearly half skip audits on subsequent upgrades. Risk concentrates in product updates: new features, dependency changes, permission modifications. A single unreviewed deployment can silently undo everything that was validated before it.
And perhaps most striking: only 50% of teams have a documented incident response plan, despite 93% actively monitoring their production environments. Detecting a breach and knowing what to do about it are two very different capabilities.
"Security is always a trade-off: perfect protection vs. time-to-market, risk appetite vs. business value. Not every control fits every context. But the gaps we identified here — regression testing, incident readiness, continuous assurance — these are the ones that hurt when ignored. I've experienced it firsthand: teams without regression tests reintroducing critical bugs, unprepared teams losing precious time while handling security incidents. The data confirms what practitioners already know: maturity isn't about doing everything. It's about knowing which risks you can take, and which ones will break you."
— Grzegorz Trawiński, Offensive Security Services Director, Hacken
Testing is another weak point. Manual and unit tests are common (90% and 77% respectively), but fuzz testing sits at just 39% and invariant testing at 32%. These advanced methods are specifically designed to catch adversarial behaviors and edge cases that standard tests miss — exactly the scenarios that get exploited in production. Meanwhile, 1 in 5 teams still doesn't run regression tests at all, meaning every hotfix carries the risk of quietly reopening a vulnerability that was already patched.
And with 80% of respondents now using AI tools in their development process, a new risk surface is growing faster than the guidance to manage it: prompt leakage, code provenance questions, and security reviews being quietly bypassed.
Who made this possible
This report reflects ground-up insights from builders across the digital assets ecosystem who chose to share their practices openly. The following organizations opted in to be publicly recognized as contributors:
Their openness is what makes industry-wide benchmarking possible.
Where to focus next
Teams that want to raise their maturity quickly should prioritize: incident response planning and tabletop exercises; formalizing security requirements and threat modeling before writing code; mandatory technical security training for developers; enforcing least privilege access with periodic reviews; automated regression testing for every update; adopting fuzz and invariant testing for critical code paths; and moving from one-off audits to release-driven assurance on every significant change.
"The key takeaway confirms Hacken's direction for 2026: moving from limited point-in-time security to continuous assurance, supported by structured training, sufficient testing, and cybersecurity embedded across the entire software lifecycle."
— Yev Broshevan, CEO & Co-Founder, Hacken
Read the full report or add your voice to the 2026 data.
The complete findings, including domain-by-domain breakdowns and maturity benchmarks, are available to download here:Download the Full Report
Hacken's 2025 SSDLC Maturity Survey
1 in 2 teams has no incident response plan. 45% ship upgrades without an audit. See how your peers are building — and where they're falling short.
We're now collecting data for the 2026 edition. The survey takes minutes, your insights inform public research used across the industry, and you can opt in to have your company featured in the final report.Take the Survey →
2026 Hacken’s 2025 SSDLC Maturity Survey
What We Learned from Hacken’s 2025 SSDLC Maturity Survey
Secure software delivery is no longer a “nice to have” — it’s a prerequisite for building reliable products and maintaining user trust. In 2025, Hacken ran a Secure Software Development Life Cycle (SSDLC) maturity survey across our customers and partners to better understand how teams build software today, where security practices are strong, and where organizations still need practical support.
This article summarizes the key signals from the survey and highlights the most actionable improvement areas we observed.
“This survey reflects what I've seen from both sides—as a developer and as a security engineer. Security is always a trade-off: perfect protection vs. time-to-market, risk appetite vs. business value. Not every control fits every context. But the gaps we identified here—regression testing, incident readiness, continuous assurance—these are the ones that hurt when ignored. I've experienced it firsthand: teams without regression tests reintroducing critical bugs, unprepared teams losing precious time while handling security incidents. The data confirms what practitioners already know: maturity isn't about doing everything. It's about knowing which risks you can take, and which ones will break you.”
Grzegorz TrawińskiOffensive Security Services DirectorLinkedIn
Survey scope and focus
The survey was designed to evaluate maturity across the entire software development journey, covering over 20 targeted questions across several critical domains:
Governance & Enablement: Team composition, Agile adoption, and security training.
Security by Design: Documentation, security requirements, risk analysis, and threat modeling.
Engineering Controls: CI/CD integration, peer reviews, and access management.
Verification & Testing: Manual, automated, and advanced testing methodologies (fuzzing and invariants).
Operations & Resilience: Production monitoring, maintenance, and incident response readiness.
The strong completion rate among participants suggests that the respondents — primarily from the Hacken ecosystem — are deeply engaged with security topics. Furthermore, most respondents expressed interest in being featured as contributors in our public reporting, signaling a strong community commitment to transparency and shared best practices.
Who responded: teams with ownership and process maturity
12.9%Hybrid Team
87.1%Fully in-House Team
Most respondents described their development teams as fully in-house (87.10%), with 12.90% operating hybrid models (in-house plus outsourced). This matters because secure delivery is often easier to institutionalize when ownership of architecture, code, and release practices remains internal.
On delivery methodology, 87.10% reported following Agile principles, suggesting most teams already operate in iterative cycles where security can—and should—be integrated into routine planning, definition of done, and release gates.
Security enablement is present — but not universal
Security training appears broadly adopted, but uneven in depth. Among respondents who answered the training question →
Completed both general security awareness and technical security training56.67%
Completed general security awareness only16.67%
Completed technical security training only13.33%
Reported no security training10.00%
Didn't know3.33%
This indicates many teams are investing in security capability, but a meaningful portion still lacks the baseline training needed to reduce avoidable vulnerabilities and operational risk.
Around 30% of surveyed teams lack formal technical security training — a concerning finding that directly undermines secure development practices. While general awareness training (phishing, social engineering) is valuable for reducing human error, it does not equip developers with the knowledge needed to identify and prevent code-level vulnerabilities — whether in web applications, smart contracts, or infrastructure. Without technical security training, developers cannot effectively implement threat modeling, conduct secure code reviews, or write tests that catch security-relevant edge cases. This gap is particularly acute in Web3 and DeFi, where contract vulnerabilities can result in direct financial loss.
The implication is clear: organizations cannot rely on external audits alone to compensate for a team that lacks foundational security engineering skills. Closing this training gap should be a priority for any organization serious about secure development. Technical security training — covering topics like secure coding patterns, common vulnerability classes, threat modeling, and security testing — is not a luxury; it is a prerequisite for building resilient systems. Teams that invest in this capability see measurable improvements in code quality, faster remediation cycles, and reduced audit findings.
On the positive side, teams report strong documentation habits: 93.33% confirmed they created functional and/or technical documentation. This is important because good documentation makes it significantly easier to introduce security requirements, threat models, test plans, and evidence-based release decisions.
The shift-left gap: security requirements and threat modeling
Two results stood out as “high leverage” areas for improvement—practices that, when introduced early, reduce downstream cost and prevent repeated classes of issues.
Security requirements: 70.00% said they define and document security requirements, while 26.67% do not.
Threat modeling: 65.52% reported conducting threat modeling, while 27.59% do not and 6.90% were unsure.
These numbers are meaningful because teams can often continue development while still missing a foundational step: defining “what secure means” for their system and proactively analyzing how it can fail.
Threat modeling tends to lag risk analysis (which was higher at 79.31%), but both are essential. Risk analysis answers what matters most; threat modeling answers how it breaks and how we stop it. Increasing adoption here is one of the fastest ways to improve maturity without relying solely on point-in-time external validation.
Strong baseline controls — least privilege remains the hardest
Respondents reported relatively strong adoption of several standard secure development practices:
However, least privilege access enforcement was notably lower at 61.29%, making it the weakest control in this set.
This pattern is common. Teams frequently implement MFA and CI/CD early, but access control maturity is harder: IAM sprawl, shared administrative roles, lack of periodic access reviews, and unclear ownership can all slow development progress. Yet least privilege remains one of the most effective controls for limiting the blast radius of compromised accounts, insider threats, and supply-chain issues.
Testing maturity: manual and unit tests are common; advanced methods are not
Regular Testing
Manual Testing0%90.32%100%
Unit Testing0%77.42%100%
Regression Testing0%75.86%100%
Most teams reported performing manual and unit testing. Where maturity drops sharply is in deeper, security-relevant testing approaches:
Advanced Testing
Fuzz Testing0%38.71%100%
Invariant Testing0%32.26%100%
This is one of the clearest technical gaps in the dataset. Manual and unit testing are valuable, but they often miss edge cases and adversarial behaviors that fuzzing and invariant/property-based testing are designed to catch. For teams building complex systems (especially those that manage value or safety-critical flows), these methods can materially reduce exploitability and improve confidence in upgrades.
A significant operational risk identified in the survey is that 20.69% of teams are not running regression tests. In a modern development environment where code is constantly evolving, regression testing is the primary defense against "breaking changes" — the accidental introduction of new bugs into previously stable and secure code. Without a robust regression suite, every update or hotfix carries the risk of silently re-opening old vulnerabilities or breaking critical security logic that was already validated.
1 in 5
teams does NOTrun regression tests
Without automated regression suites, new changes can break existing functionality and introduce new vulnerabilities or reopen previously resolved ones.
This gap is particularly dangerous when providing updates or patches under pressure. When developers modify existing logic to add a feature or fix a bug, they may inadvertently disrupt dependencies or edge cases they aren't currently focused on. For Web3, where code is often immutable or governs high-value assets, the absence of regression testing means that a single update could compromise the integrity of the entire system. Implementing automated regression suites is a non-negotiable step for teams moving toward a mature, resilient SSDLC.
Audits are common — but continuous assurance is not
External assurance is widely used:
70.97% reported running a security audit
25.81% have not
Regression testing: 75.86% (with 20.69% not running regression tests)
Version 1.0
Audited (safe)
29% No Initial Audit
Nearly one-third of projects enter the market without any formal security verification of Version 1.0.
Update 2.0new code added
45% Deployed without Audit
Almost half of all surveyed organizations skip security audits for subsequent code deployments.
Production
Critical Risk
Unverified updates in production create blind spots that can lead to catastrophic smart contract failures.
But when asked whether audits occur for every upgrade, responses split:
51.72% yes
44.83% no
This suggests many organizations still operate with a “point-in-time” assurance model: audit once, then iterate quickly. In practice, risk concentrates in upgrades—new features, dependency changes, configuration changes, and permission changes. This gap highlights a growing need for tiered, release-driven assurance approaches (lighter reviews for low-risk changes, deeper reviews for high-risk releases), supported by strong automated controls and testing evidence.
Operational maturity: monitoring is high, incident readiness is not
Monitoring
Actively monitor their production environment0%93.33%100%
Perform maintenance activities0%86.21%100%
Incident Response
incident response plan prepared and implemented0%50.0%100%
Two operational signals appeared together — and they tell an important story.
On one hand, 93.33% of respondents actively monitor their production environment, and 86.21% perform maintenance activities such as hotfixes or upgrades. This indicates teams are operating real systems, watching for issues, and shipping changes over time.
On the other hand, only 50.00% reported having an incident response plan prepared and implemented, while 40.00% do not and 10.00% are unsure.
High Alertness
93%Actively Monitor Production Environment
Low Readiness
Only 50%Have an Incident Response Plan
This is arguably the most critical maturity gap in the survey. Monitoring helps detect problems; incident readiness determines whether teams can respond decisively under pressure—contain impact, communicate effectively, and recover safely. In real-world security events, the absence of a practiced plan often translates into delayed decisions, unclear authority, and avoidable escalation.
AI tool adoption: a new and fast-growing risk surface
Do you make use of AI tools within the development process?
19.4%No
80.6%Yes
A large majority (80.65%) reported using AI tools within the development process.
This adoption creates clear demand for updated SSDLC guidance: handling sensitive data in prompts, preventing secret leakage, establishing code provenance expectations, and ensuring AI-assisted coding does not bypass secure review and testing standards. AI can accelerate development, but it also increases the importance of guardrails, traceability, and strong “definition of done” criteria.
What these results suggest (and where teams can act next)
Across the survey, teams show strong momentum: documentation, Agile delivery, CI/CD, reviews, MFA, and production monitoring are widely present. The highest-value improvements are concentrated in a set of practices that raise maturity quickly:
01
Incident response planning and tabletop exercisesMoving beyond monitoring to ensure the team can contain, communicate, and recover from exploits under pressure.
02
Formalizing security requirements and threat modelingShifting security to the design phase to identify "how it breaks" before a single line of code is written.
03
Mandatory technical security trainingClosing the 30% gap by equipping developers with the specific skills needed to prevent code-level vulnerabilities and smart contract exploits.
04
Enforcing least privilege and periodic access reviewsHardening the development environment by ensuring access is granular, monitored, and revoked when no longer needed.
05
Automated regression testing for every updateEnsuring that new features or hotfixes do not silently re-introduce old bugs or break existing security logic in stable code.
06
Adopting advanced testing (fuzzing and invariants)Moving beyond basic unit tests to catch complex edge cases and adversarial behaviors in critical code paths.
07
Release-driven assurance for upgrades (continuous assurance)Transitioning from "one-off" audits to a model where every significant change is validated before deployment.
Achieving these milestones is not about reaching a state of 'perfect security'—an impossible standard in a shifting threat landscape. Instead, it is about moving from ad-hoc reactions to a model of consistent habits, clear ownership, and repeatable workflows that make security a natural part of the software engineering process.
“Hacken’s SSDLC Maturity Survey provides a transparent benchmark for everyone shaping secure software, including development teams, product owners, protocol founders, and investors offering clarity on current market practices and where gaps remain.
For me, the key takeaway confirms Hacken’s direction for 2026: moving from limited point-in-time security to continuous assurance, supported by structured training, sufficient testing, and cybersecurity embedded across the entire software lifecycle.”
The goal of Hacken’s SSDLC maturity survey was simple: understand how teams actually build software today, identify gaps, and use that insight to prioritize practical improvements. Even with a limited sample size, the patterns are consistent with what we see across real engagements: many organizations have built solid foundations, but still need help operationalizing security early and making it sustainable through upgrades and incidents.
Contribute to the 2026 SSDLC Survey
We’re mapping how secure software is built in the new digital frontier.
Share your perspective and help define the benchmark. Participants can opt to be featured in the public report.
The following organizations agreed to be publicly acknowledged as contributors to the 2025 SSDLC Maturity Report.
Security is a shared responsibility
The 2025 SSDLC Maturity Survey is an industry-wide research initiative led by Hacken in collaboration with builders, founders, and engineering leaders from across the digital assets ecosystem: from Layer 1 infrastructure and DeFi protocols to payment rails, cross-chain tooling, and community-owned financial tools. This report reflects ground-up insights by the teams actively navigating secure software development in their day-to-day work. By pooling knowledge across organizations of different sizes, structures, and maturity levels, this effort aims to give the industry an honest, data-driven benchmark.
The contributors listed in this report voluntarily chose to be publicly recognized for their participation. Their openness is what makes cross-industry benchmarking possible, and what helps the entire ecosystem build more securely.
Making Web3 a safer place
Hacken is an end-to-end blockchain security & compliance partner for digital assets
Hacken Track Record
9+years securing blockchain and digital assets
2,168public security assessments delivered since 2017
16,735vulnerabilities identified across engagements
$430B+in assets verified through Proof of Reserves services
ISO 27001certified information security management
Unlike traditional providers, Hacken was born on blockchain, combining deep Web3 expertise with enterprise-grade quality, AI-powered offensive security, and globally recognized certifications. Since 2017, Hacken has been trusted by startups, enterprises, and regulators to secure the new digital frontier.