HashiCorp Vault CVE-2026-89322

Exploitation requires an authenticated Vault principal that has been granted delegated permissions capable of referencing resources or assigning policies, operating against an ACL policy set that relies on an explicit deny rule to protect a resource or restrict a policy. The flaw triggers only when a resource or policy name is supplied in a non-canonical form that the ACL engine evaluates inconsistently versus its canonical form (CWE-178). Limiting factors: the attacker cannot be anonymous or unauthenticated (PR:H), must already possess delegated management rights, and the target environment must depend on explicit deny restrictions - single-operator or allow-only policy models with no delegated sub-privileges present little exploitable surface. No user interaction by a victim is required (UI:N).

This is a credible but moderate-priority issue, best treated as a defense-in-depth hardening fix rather than an emergency. The CVSS 7.2 (High) vector CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H combines a network attack vector and low complexity with a critical gating factor: PR:H, meaning the attacker must already be an authenticated user holding delegated (policy/privilege-management) permissions. The full C:H/I:H/A:H triad reflects that successful privilege escalation can grant access to arbitrary protected secrets and policies. No EPSS score was provided in the input, and there is no CISA KEV listing and no public exploit identified at time of analysis, so there is no evidence of active or at-scale exploitation - risk is currently theoretical and insider/post-compromise oriented. The key tension to note: CVSS AV:N implies remote network reach, but the flaw is fundamentally an authorization-logic bug (CWE-178, inconsistent canonicalization of resource/policy names) reachable only by an already-trusted, authenticated principal with delegated rights - so the realistic threat model is a lower-privileged-but-authorized tenant escalating against an explicit deny, not an anonymous remote attacker. Prioritize patching in multi-tenant or delegated-administration Vault deployments where many principals hold policy-assignment rights; the risk is materially lower in single-operator or tightly-scoped deployments.

An authenticated Vault user who has been delegated limited policy- or resource-management permissions crafts resource or policy names in a non-canonical form (e.g. differing case or encoding) that bypass an explicit deny rule the ACL engine fails to apply against the canonical name. The user thereby reads a protected secret path they were denied, or assigns themselves a policy that was explicitly denied, escalating their effective privileges. No public exploit code was identified at time of analysis, and exploitation requires pre-existing authenticated access with delegated permissions.