GCP public images show how curation can reduce secret exposure

At a glance

What this is: Truffle Security found no exposed secrets across 8,437 public GCP images, suggesting that tighter image curation materially reduces credential leakage risk.

Why it matters: IAM, PAM and NHI teams should read this as evidence that publication controls can suppress one exposure path, but secrets still need lifecycle governance outside the image marketplace.

By the numbers:

• Truffle Security processed 8,437 public GCP images and found zero exposed secrets.
• The researchers extracted more than 100 GB of potentially sensitive files from the scanned images.

👉 Read Truffle Security's analysis of secret exposure in public GCP images

Context

Public cloud images can become a secret exposure path when credentials, tokens or configuration files are baked into reusable machine artifacts. That risk is especially relevant to IAM and NHI governance because the same leaked credential can unlock services far beyond the original image.

Truffle Security’s finding matters because it separates the image publication problem from the broader secrets management problem. GCP’s curated marketplace appears to reduce exposure at the distribution layer, but teams still need controls for creation, storage, rotation and offboarding across the rest of the estate.

Key questions

Q: What fails when public cloud images are published without secret screening?

A: The failure is that one embedded credential can be copied into many deployments before anyone notices. Once a secret is baked into an image, publication can turn a local build mistake into a multi-environment access problem. The control gap is pre-publication scanning and release approval, not just post-release incident response.

Q: Why does image curation reduce secret exposure risk?

A: Curated publication adds a gate between image creation and public release, so bad artifacts are less likely to reach consumers. That lowers exposure probability, but it does not protect against every secrets pathway in the environment. Teams still need rotation, revocation and inventory controls for credentials outside the image lifecycle.

Q: How do security teams know whether public image controls are actually working?

A: They should look for whether secret scanning happens before publication, whether exceptions are tracked, and whether leaked credentials are rotated quickly when found. A low or zero finding rate is useful, but only if the pipeline also proves that scanning coverage is broad enough to be trusted.

Q: Should organisations rely on curated marketplaces instead of secrets governance?

A: No. Curated marketplaces can reduce one source of accidental exposure, but they do not replace credential lifecycle management across builds, registries and runtime systems. The right model is layered control, where curation, scanning, rotation and workload identity each address a different part of the exposure chain.

Technical breakdown

Why public cloud images leak secrets

Public images are snapshots of software and system state, so any embedded token, key, config file or shell history can travel with the image into every future deployment. The risk is not limited to human logins. Service credentials, API keys and cloud access tokens frequently appear in build artifacts, startup scripts and application directories. Once an image is shared broadly, the exposure becomes durable because the same secret may be replicated into multiple instances, environments and accounts.

Practical implication: treat public-image publication as a secrets boundary and scan image contents before release.

How curation changes the exposure model

GCP’s public-image model is curated rather than open publication. Approved publishers and marketplace vendors act as a gate before an image becomes public, which creates an additional control point for screening and validation. That does not prove perfect detection, but it materially changes the probability of accidental credential exposure compared with open publishing models where any user can expose an image. The control is therefore about preventing bad artifacts from entering circulation, not fixing secrets after they are already distributed.

Practical implication: place validation controls upstream of publication, not only at the point of secret discovery.

Why zero findings still matter for governance

A zero-finding scan does not mean the risk has disappeared. It means one distribution channel is better controlled than others, which shifts attention to development pipelines, internal images, marketplace exceptions and paid images that were not included in the study. For NHI governance, the lesson is that short-lived secrets and workload identity reduce reliance on image hygiene, but they do not remove the need to inventory credentials wherever machines are built or cloned.

Practical implication: govern image publication and secrets lifecycle as separate but connected controls.

Threat narrative

Attacker objective: The attacker’s objective is to recover a reusable credential from a public image and use it to access downstream cloud, application or SaaS resources.

1. Entry occurs when a secret is embedded in a machine image during build or packaging.
2. Credential abuse would follow if that image is published publicly and reused in deployment.
3. Impact would be credential compromise across environments that inherit the same baked-in secret.

Breaches seen in the wild

• Hugging Face Spaces breach 2024: Unauthorised access to Hugging Face Spaces may have exposed secrets users stored for AI apps; tokens were revoked and org tokens removed.
• Red Hat Consulting GitLab breach 2025: Extortionists copied a Red Hat Consulting GitLab instance, exposing customer tokens, keys and database connection strings held in engagement reports.

Read our 52 NHI Breaches Analysis report for a comprehensive view of breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Curated publication is an exposure-control layer, not a secrets strategy. Truffle Security’s result shows that tighter marketplace governance can materially lower the chance of leaking credentials through public images. But that protection only applies at one point in the lifecycle, so teams still need scanning, rotation and revocation outside the publishing gate. Practitioners should treat curation as a risk reducer, not a substitute for secrets governance.

Image-based secret leakage is a governance failure of reuse, not just a scanning failure. The real problem is that one credential embedded in a build artifact can inherit the blast radius of every future deployment. That is an NHI issue as much as a cloud issue, because service accounts, API keys and tokens often outlive the image they were placed in. Practitioners should separate artifact hygiene from credential lifecycle management.

GCP’s marketplace model suggests a named concept: public image curation as a secret suppression control. The article’s core signal is that a controlled publication pipeline can suppress accidental exposure before it becomes an incident. That does not eliminate the need for runtime and post-publication detection, but it does show that pre-publication screening has measurable governance value. Practitioners should ask where else curated release gates can reduce repeatable secret leakage.

The absence of findings should prompt better exception handling, not complacency. Zero exposed secrets in one cloud provider’s public images does not generalise to internal images, private registries or other platforms with looser controls. The lesson for security leaders is to map which image channels are curated, which are merely reviewed, and which are effectively open. Practitioners should base policy on channel risk, not on a single positive result.

This research reinforces the separation between distribution controls and credential controls. A secure marketplace can reduce exposure at the point of publication, but it cannot fix long-lived secrets, reused tokens or unmanaged service accounts elsewhere in the estate. That makes image curation a useful layer in a broader identity governance model, not a replacement for it. Practitioners should align cloud publishing controls with secrets lifecycle ownership.

What this signals

Public-image curation is becoming a governance pattern, not just a marketplace feature. The article shows that controlling who can publish images can materially change the risk profile of exposed secrets. For practitioners, that means the publishing gate deserves the same design attention as the image scanner itself.

Secrets management and cloud publishing need to be governed as one chain. If a service account, token or API key can end up inside an image, the publication process becomes part of credential risk management. Teams should make sure their NHI inventory, rotation policy and release controls are aligned rather than operated as separate programmes.

For practitioners

• Scan public-image pipelines before publication Add secret detection to the image build and release path so credentials are caught before an image can be made public or shared internally.
• Separate image curation from credential governance Track which teams own image publication approvals and which teams own secret rotation, revocation and offboarding so the controls do not get confused.
• Review exceptions for paid and private images Extend the same checks to marketplace exceptions, licensed images and private registries, because the study excluded those channels and they may carry different exposure patterns.
• Reduce reliance on baked-in secrets Replace embedded API keys and service credentials with workload identity or short-lived tokens wherever the deployment model allows it.
• Map public-image governance to NHI ownership Assign clear ownership for service accounts, tokens and certificates that could be duplicated into images, including rotation and revocation responsibilities.

Key takeaways

• Public GCP images showed no exposed secrets in this scan, which suggests that curated publication can suppress one important leakage path.
• The result does not remove the underlying problem of baked-in credentials, because secret governance still has to cover builds, registries and runtime systems.
• Security teams should treat image curation as one control layer and continue to enforce inventory, rotation and revocation for all machine credentials.

Key terms

• Public Image Curation: Public image curation is the process of restricting who can publish machine images and screening them before release. In cloud governance, it acts as a release gate that can lower accidental secret exposure, but it does not replace credential lifecycle controls elsewhere in the environment.
• Baked-In Secret: A baked-in secret is a credential, token or key embedded in an image, build artifact or configuration file before distribution. Once released, it can be replicated across many systems, which makes the original mistake much harder to contain than a secret kept in a managed vault.
• Secrets Lifecycle: Secrets lifecycle is the management of credentials from issuance through rotation, revocation, and offboarding. It matters because a secret that is technically valid can still be operationally unsafe if its owner, purpose, or downstream access paths are no longer current.
• Image Publication Gate: An image publication gate is a control that blocks unreviewed or non-compliant machine images from becoming publicly available. It can reduce exposure by forcing validation before distribution, but its value depends on how well it is linked to scanning, ownership and exception handling.

What's in the full report

Truffle Security's full post covers the operational detail this post intentionally leaves for the source:

• The image discovery and extraction workflow used to process 8,437 public GCP images
• The file-extension targeting logic used to narrow extraction to likely secret-bearing files
• The comparison with prior AWS AMI and Azure image research across cloud platforms
• The limitations of excluding paid images and why that matters for future research

👉 The full Truffle Security post covers the scanning method, platform comparison and research limitations in detail.

Deepen your knowledge

The NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, machine identity security and secrets management. It helps practitioners connect identity controls to broader security programmes that need consistent ownership and lifecycle discipline.