Summary
@hulumi/policies contains a parent-spoofing weakness that allows an attacker to submit falsified SecureBucket parent evidence during policy evaluation. When the validator trusts this evidence, unsafe bucket configurations may pass security policy checks.
Affected context
The vulnerability occurs in deployments using the affected @hulumi/policies policy evaluation logic for SecureBucket resources. Risk is present when an untrusted or insufficiently authenticated caller can provide parent relationship evidence that the validator accepts without independently confirming the resource hierarchy.
Exploit preconditions
- Attack surface: Policy evaluation or validation interfaces that process SecureBucket parent evidence
- Required access: Ability to submit or influence evidence consumed by the validator
- Trigger point: Evaluation of a SecureBucket configuration and its claimed parent relationship
- Why exploitation is possible: The validator may trust caller-provided parent evidence instead of verifying the relationship against an authoritative resource source
Why this happens
The underlying bug is an authorization-boundary failure in which security decisions rely on caller-controlled relationship metadata. By accepting spoofed parent evidence, the policy engine can associate a resource with an apparently trusted parent and skip checks that should apply to the actual configuration. This behavior maps to CWE-284: Improper Access Control, because the validator does not reliably enforce authorization and policy constraints using trusted resource relationships.
Potential impact
- Attackers may bypass security policy checks
- Unsafe SecureBucket configurations could be accepted
- Misconfigured storage resources may be deployed or retained
- Organizations could experience unauthorized exposure of bucket contents or metadata
- Downstream automation may propagate invalid compliance decisions
General mitigation guidance
Mitigation generally focuses on making authorization and relationship validation independent of caller-supplied evidence.
Common defensive patterns include:
- Resolve parent-child relationships from authoritative state rather than submitted claims
- Treat policy evidence as untrusted input and validate its provenance
- Apply deny-by-default behavior when resource ownership or hierarchy cannot be confirmed
- Separate identity, resource ownership, and policy evaluation concerns
- Run policy evaluation with only the permissions required to inspect relevant resources
- Add regression tests for parent substitution, mismatched identifiers, and cross-tenant relationships
How teams detect issues like this
Organizations can look for this class of issue through:
- Static analysis and code review checks for authorization decisions based on caller-controlled relationship fields
- Review signals where resource parent identifiers are accepted without authoritative lookup
- Policy tests that substitute valid parent evidence from an unrelated resource
- Runtime logging of evaluated resource identifiers, resolved parents, and evidence provenance
- Enforcement rules requiring trusted resource resolution before security-sensitive decisions
- Automated security review tools that identify missing authorization checks and trust-boundary violations

