Safeguard
Security Guides•

How do you find the owner of a vulnerable asset with no tags?

Follow the trail from image annotations to repository, commit history and account owner. A lookup order for untagged assets, and how to stop it recurring.

Safeguard Security Team
3 min read

To find the owner of an untagged vulnerable asset, follow the build trail: image source and revision annotations, then the repository and its commit history, then the account or site owner. Record whatever you find as a tag straight away, so the next finding routes itself.

The lookup order

Go from the most reliable evidence to the least.

  1. Image annotations. The OCI spec defines org.opencontainers.image.source for the URL of the source and org.opencontainers.image.revision for the revision. If present, you have the repository and commit.
  2. Repository ownership. Read the repository's owners file or team settings, and look at who merged recent changes.
  3. Deployment records. The pipeline that deployed the asset, or the infrastructure code that defines it, often names a team.
  4. Naming conventions. Names may carry a team prefix. Treat that as a lead, not proof.
  5. Account or site owner. The owner of the account is accountable for what runs in it, even if they did not build it.
  6. Network and access clues. Which systems talk to it, and which identities have deploy access.

Record what you find

EvidenceConfidenceNext step
Source annotation plus owners fileHighAssign directly
Deployment pipeline ownerMediumConfirm with the team
Name prefix onlyLowAsk the likely team
Account owner onlyFallbackAsk them to find the builder

Always ask the likely owner to confirm. A wrong assignment wastes days.

When nobody claims it

If no one admits ownership, do not close the finding. Options in order:

  • Assign it to the account owner with a deadline for finding the builder.
  • If the asset is exposed to the internet and nobody can justify it, plan to isolate or remove it, following your change process.
  • Add it to a visible list of orphans, reviewed in the weekly meeting.

Stop it happening again

  1. Require tags in templates. The AWS tagging guidance describes tags as the mechanism to track which resources are assigned to which application, including for patching and incident management.
  2. Add annotations to builds. Make source and revision part of the shared pipeline template.
  3. Report coverage. Track the share of assets with an owner and drive it up.
  4. Block new untagged assets where the platform allows policy checks.

An example trail

A scan flags an old image with no tags. Its source annotation points to a repository that was archived last year, but the owners file still lists the data engineering team, and the deployment pipeline posts to their channel. A message to that channel gets a reply the same day: the job is retired, nobody switched it off. The image is isolated, then removed through the normal change process, and the finding closes with a note.

Quick check

Is the person who last committed the owner?

Not necessarily. They are a lead. Confirm with the team and record the team, not the person.

Can I delete an asset nobody owns?

Follow your change process. Isolating it first and watching for complaints is safer than immediate deletion.

Safeguard's asset discovery can enumerate what runs, and a free scan is available at gold.safeguard.sh/scan.

Sources

Last reviewed: 19 October 2026.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.