804966
1656687

The Credential Layer Is Expanding Faster Than Security Teams Can See

The Credential Layer Is Expanding Faster Than Security Teams Can See

The Credential Layer Is Expanding Faster Than Security Teams Can See

Modern organisations rely on credentials to connect people, applications, cloud platforms, infrastructure and increasingly AI-powered systems. Every password, API key, authentication token and other secret can provide legitimate access to important resources, but each one can also become a security risk if it is exposed or poorly managed.

The challenge is becoming more complicated as software development accelerates. More code means more applications, integrations, automated workflows and services. As those environments expand, the number of credentials required to connect them also increases.

This creates a growing credential layer that can extend across repositories, cloud environments, employee devices, collaboration platforms and AI tools. Security teams therefore face a difficult question: how can they protect credentials they cannot fully see?

Recent developments in software production demonstrate the scale of the problem. GitHub has experienced enormous growth in the volume of code being produced, with the platform reportedly moving from about one billion commits during 2025 to a pace of billions of commits in 2026.

541599 (2)

The expansion is not simply about the amount of code being written. Every new application and service can require additional authentication. New integrations create new connections, while automation and AI coding agents introduce additional systems that may need access to data and infrastructure.

Research cited by GitGuardian found that 28.65 million new hardcoded secrets were detected in public GitHub commits during 2025, representing a 34 percent increase from the previous year. The same research indicated that credentials associated with AI services increased significantly.

These figures demonstrate why organisations need greater visibility into the credentials being created, used and distributed throughout their environments.

The credential layer has no clear boundary

Unlike a traditional network perimeter, the credential layer does not stop at the edge of an organisation‘s infrastructure.

A developer may create a credential inside an approved cloud environment, use it during development and later copy it into a source-code repository, collaboration platform or local configuration file.

Another credential may remain inside an approved secrets-management system while an unprotected copy sits on a developer’s computer.

Credentials can also emerge outside established security processes when employees adopt new tools, experiment with personal projects or use AI-powered development services.

This creates an attack surface that crosses organisational, technological and ownership boundaries.

Security teams may have excellent visibility into credentials stored in approved vaults while having little awareness of secrets residing elsewhere.

That creates a significant discovery problem.

A repository scanner can identify credentials inside source code. An endpoint security tool can reveal information stored on a workstation. A secrets-management platform can provide information about credentials under central control.

But none of these individual systems necessarily provides a complete picture.

Security teams need to connect those perspectives to understand the full credential population.

Every discovery source reveals only part of the picture

Source-code repositories remain one of the most important places to search for exposed credentials.

A secret may be removed from the latest version of a file but remain inside the repository’s history. Copies can also exist in branches, forks or other repositories.

Once credentials have been publicly exposed, controlling their distribution becomes significantly more difficult.

Internal repositories create another challenge. They can contain credentials used by developers and applications to access internal services, cloud platforms and production environments.

Collaboration tools create yet another exposure point.

During troubleshooting, employees may paste credentials into tickets or messages. Temporary information shared during a technical handover can remain searchable long after the original problem has been resolved.

The result is a credential ecosystem that extends well beyond source code.

Developer devices are becoming critical credential repositories

Developer computers have always been important because software development requires access to multiple systems.

Developers may use local configuration files, command-line tools, cloud authentication systems and application credentials during their work.

However, the security significance of these devices has increased as attackers have increasingly targeted them.

Infostealer campaigns have demonstrated how valuable developer endpoints can be. A compromised device may contain authentication material capable of providing access to source-code repositories, cloud infrastructure, development environments and other enterprise resources.

This means a developer laptop should not be treated simply as an endpoint. It can also represent a concentrated collection of credentials connecting multiple parts of an organisation.

Repository scanning can reveal credentials that entered source control, while endpoint discovery can identify secrets that never reached a centrally monitored repository.

Public monitoring can provide another perspective by identifying credentials exposed outside corporate systems.

The credential layer becomes much clearer when all of these sources are considered together.

Attackers are already exploiting these boundaries

Threat actors have increasingly recognised that credentials can provide a shortcut into enterprise environments.

According to data cited from the 2026 Verizon Data Breach Investigations Report, compromised credentials accounted for a significant portion of initial access incidents.

The problem is that attackers do not necessarily care whether a credential was created by an employee, an application or an automated system.

If a valid credential provides access, it can potentially become useful regardless of where it originated.

Infostealers can search local files, browser information and application storage for authentication material. They can collect credentials without regard for the internal ownership structure that security teams use to organise their environments.

A compromised developer device can therefore become a bridge between several systems that might otherwise appear to be separate.

A single machine could have access to source control, cloud services, development environments and third-party applications.

Security teams need to understand those connections before attackers discover them.

AI agents are creating another layer of complexity

The modern developer environment is changing again with the rapid adoption of AI coding agents.

These systems can read files, execute commands and interact with external tools and services. Model Context Protocol connections and similar integrations can expand what AI systems are capable of accessing.

Every additional connection creates another point where authentication and authorisation must be properly controlled.

This does not mean AI agents are inherently unsafe. The concern is that organisations may not always have complete visibility into what credentials these systems can access or what actions they are authorised to perform.

If an AI agent has access to a development environment containing sensitive credentials, an attacker who compromises that environment may potentially gain access to those same resources.

There is also the possibility of unintended actions.

An automated system operating with excessive permissions could make changes that were never intended by its users.

Security teams therefore need to understand exactly what AI agents can access, what credentials they use and what permissions those credentials provide.

Discovery alone is not enough

Finding a credential is only the beginning.

Security teams also need information that explains how dangerous a particular credential might be.

One important factor is whether the credential is still valid.

An exposed credential that has already expired presents a different risk from one that remains active.

Research cited by GitGuardian found that many credentials previously confirmed as valid remained valid years later, highlighting the importance of quickly identifying and revoking exposed secrets.

Location is another important factor.

A credential found inside an internal repository has a different exposure history from one that appears on a developer laptop and later becomes publicly accessible.

Tracking where the same credential appears can help security teams understand how far it has travelled.

Ownership is equally important.

Security teams need to know which person, application, service or workload is responsible for a credential.

A finding without an owner can become difficult to remediate because nobody is clearly responsible for rotating or revoking it.

Permissions provide another critical piece of information.

A credential that can access a limited development environment presents a different risk from a credential with broad administrative privileges across production systems.

The more powerful the credential, the more urgent its protection becomes.

Dependencies also matter.

Before revoking a credential, organisations need to understand which applications or services depend on it. Removing a credential without understanding its dependencies can disrupt legitimate operations.

The ideal inventory therefore does more than identify secrets. It connects each credential to its validity, location, owner, permissions and dependencies.

Security teams need to know what they cannot see

Many organisations already have sophisticated secrets-management programmes.

Approved vaults and authentication platforms can provide strong controls over credentials that are properly registered and managed.

The problem begins when organisations assume that the credentials inside those systems represent the entire credential population.

They may not.

A company could have tens of thousands of credentials inside approved systems while additional secrets remain in source-code repositories, developer devices, collaboration platforms or third-party services.

Some of these may be duplicates of credentials already managed centrally. Others may be completely unknown to the security team.

Without discovering the broader population, organisations cannot accurately measure their coverage.

Security leaders need to know how many credentials have identified owners, how many remain active, which ones have privileged access and which have been exposed publicly.

These measurements are only possible when the organisation has first established a comprehensive inventory.

Software development is accelerating

The pressure on security teams is increasing because software development itself is accelerating.

More applications and more frequent deployments mean more credentials and integrations.

The growth of AI-assisted development could accelerate this even further as coding agents increase the speed at which software is created and modified.

This creates a difficult imbalance.

If credentials are being created faster than security teams can discover and govern them, the visibility gap will continue to grow.

Attackers, meanwhile, are also becoming faster.

Threat research has documented increasingly short periods between initial compromise and subsequent attacker activity.

This means organisations cannot rely entirely on slow, manual discovery processes.

Security teams need visibility that operates at the same speed as modern development environments.

Detection must come before remediation

The first step in reducing credential risk is knowing what exists.

Organisations need to discover credentials across source-code repositories, public platforms, employee endpoints, collaboration systems and other environments where secrets may appear.

They then need to enrich those discoveries with information about validity, ownership, permissions and dependencies.

Once the credential population is understood, security teams can begin prioritising remediation.

High-risk credentials can be revoked or rotated. Publicly exposed secrets can be addressed urgently. Unnecessary credentials can be removed, while legitimate credentials can be placed under stronger management.

Prevention can then focus on stopping the same exposure patterns from happening again.

This creates a continuous cycle of detection, remediation and prevention.

The future credential layer will be even larger

The expansion of software development is unlikely to slow down.

More applications, cloud services, automated workflows, APIs and AI agents will continue to create connections between systems.

Every connection can require some form of authentication.

That means the credential layer will continue expanding alongside the technology environment.

Security teams therefore need to move beyond the assumption that credentials exist only inside approved vaults or source-code repositories.

They must consider the entire ecosystem in which credentials are created, copied, stored, used and exposed.

The goal is not simply to discover more secrets.

It is to understand the relationships surrounding them.

A credential is valuable to an attacker because of what it can unlock. Knowing where it exists, whether it remains valid, who owns it, what permissions it has and which systems depend on it gives defenders the context needed to make better decisions.

The growing use of AI adds another dimension to this challenge.

As organisations give automated systems greater access to development tools and enterprise resources, security teams must ensure that those systems operate with appropriate permissions and that their credentials are properly monitored.

The fundamental principle remains the same: you cannot protect what you cannot see.

As software production continues to accelerate, the organisations best positioned to manage credential risk will be those capable of maintaining visibility across the entire credential layer rather than focusing only on the areas already covered by existing security controls.

Detection provides the map.

Remediation removes unnecessary risk.

Prevention helps stop the same problems from returning.

Together, these measures can help organisations keep pace with an expanding digital environment where humans, applications and AI systems increasingly depend on credentials to communicate and operate.

The challenge for security teams is no longer simply protecting passwords and API keys.

It is understanding the growing network of trust that those credentials create — and ensuring that attackers do not discover its weaknesses before defenders do.

If you found this article useful, follow us on Google News, X and LinkedIn for more cybersecurity news, security updates, vulnerability reports and exclusive technology content.

1 thought on “The Credential Layer Is Expanding Faster Than Security Teams Can See”

  1. Pingback: GhostAction Cyberattack Hits 1,900 GitHub Accounts, Targets More Than 30,000 Repositories - GEENCEE

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top