English
Español
Français
Deutsch
日本語
한국어
العربيةSelf-grant ECR pull/push via repository policy
Can rewrite an ECR repository's policy to grant itself pull/push, then pull or poison images.
Grants escalation on its own.
ecr:SetRepositoryPolicyOn its own, ecr:SetRepositoryPolicy can rewrite an ECR repository's policy to grant itself pull/push, then pull or poison images. On a wide resource scope that alone reaches access for an outside or public principal.
Why it is dangerous
A principal can call ecr:SetRepositoryPolicy on the repository to attach a policy granting itself (or an external account / the public) pull or push. Pulling reads container images, which frequently bake in credentials, tokens and source - direct data exfiltration. Pushing replaces image tags a downstream ECS/EKS/CodeBuild consumer trusts, poisoning the supply chain so the attacker's code later runs as the consumer's role. The attacker rewrites the lock rather than holding a key.
In MITRE ATT&CK terms this maps to T1098, Account Manipulation, under the privilege escalation tactic. AWS records it in CloudTrail as SetRepositoryPolicy. The call shows up in CloudTrail but blends into normal administration, so it needs a rule to stand out.
Detection
SetRepositoryPolicyPrerequisites
ecr:SetRepositoryPolicy on the repository. The abuse additionally assumes no SCP/boundary/PrincipalOrgID guardrail blocks the over-grant; the push variant assumes a privileged consumer later pulls the poisoned tag.
MITRE ATT&CK
T1098 · Account ManipulationTactic
Privilege Escalation
A control that would remove it: Restrict ecr:SetRepositoryPolicy to registry admins; deny repository-policy writes via SCP; require aws:PrincipalOrgID on repo policies; enable tag immutability + image scanning so a poisoned push is caught.
Whether to apply it depends on how the permission is used in your environment, so it is a business decision rather than an automatic fix. The same permission can be load-bearing in one account and pure risk in another.
Pathbreak confirms whether this path resolves in your account, ranks it against the others by how much it opens up, and uses your threat model and business context to show which paths are worth cutting.
Questions
What does ecr:SetRepositoryPolicy allow?
Can rewrite an ECR repository's policy to grant itself pull/push, then pull or poison images.
Can ecr:SetRepositoryPolicy escalate privileges on its own?
Yes. Held on a broad enough resource it reaches access for an outside or public principal with no other permission.
How is self-grant ECR pull/push via repository policy detected?
Watch for SetRepositoryPolicy in CloudTrail. The call shows up in CloudTrail but blends into normal administration, so it needs a rule to stand out.
How do you break this path?
A control that would remove it is: Restrict ecr:SetRepositoryPolicy to registry admins; deny repository-policy writes via SCP; require aws:PrincipalOrgID on repo policies; enable tag immutability + image scanning so a poisoned push is caught. Whether to apply it depends on how the permission is used in your account, so it is a business decision rather than an automatic fix.
Which MITRE ATT&CK technique is this?
T1098, Account Manipulation, under the privilege escalation tactic.
Related paths
References
Find this in your own account
We run a read-only scan of one AWS account and show you which of these paths exist, ranked by blast radius, with the change that breaks each one. First validated paths within the hour.