When a JWT cannot be revoked, access control becomes stale the moment a user’s rights change. A token issued earlier can still authorize actions that the server no longer intends to allow, creating the New Enemy Problem. That gap is especially dangerous for backend services, where revoked users may still reach sensitive data until the token expires.
Why a JWT becomes unsafe after permissions change
A JWT is self-contained, so the server usually trusts the claims it already issued until the token expires. If permissions change but the token cannot be revoked, the token remains a valid bearer of the old authority. That breaks the assumption that authorization state is current at request time, not just at issuance time.
In practice, the failure is not only “old access continues”, but “policy and token state diverge”. The JWT Security Guide covers the controls that reduce that gap, including strict validation, short lifetimes, and revocation-aware design.
This is why JWTs work best when you can tolerate short-lived stale authorization, or when you pair them with a mechanism that can re-check policy before sensitive actions. Without that, the token becomes a temporary exception to the latest access decision.
What the New Enemy Problem means in token-based access
The New Enemy Problem is the mismatch between a user’s current rights and a previously issued token that still represents old rights. That mismatch matters most when permissions are reduced, a role is removed, or an account is disabled but the token is still accepted by downstream services. The problem is common in distributed systems because many services validate signature and expiry, but not the current entitlement state.
Backend APIs and service-to-service paths are especially exposed because they often consume access tokens without an interactive user in the loop. The Token and Session Security Guide is useful here because it treats revocation, replay, and lifetime as part of one control problem rather than separate concerns.
Once a token is minted, the system has to choose between stateless convenience and immediate revocation. If the design favours statelessness too heavily, the security boundary shifts from “who is allowed now” to “who was allowed when the token was issued”.
Where this creates the most operational exposure
The greatest exposure appears when permissions are removed for a reason that should take effect immediately, such as offboarding, role removal, privilege reduction, incident response, or compromise containment. In those cases, a still-valid JWT can keep authorising access to data or actions that the business has already decided to remove. That is especially dangerous for privileged users, integrations, and backend services with broad downstream reach.
For machine-facing systems, the issue often extends beyond one token. A long-lived credential chain, cached authorisation decision, or reused token pattern can keep multiple systems aligned to stale privilege. The Guide to SPIFFE and SPIRE is a useful comparison point because it shows how workload identity, attestation, and trust bundles can support tighter trust boundaries than a purely bearer-token model.
When token revocation is unavailable, the practical question becomes how quickly the system can narrow the blast radius. The shorter the token lifetime and the more sensitive the operation, the less tolerance there is for delayed policy propagation.
Risk and Threat Considerations
The main risk is stale authorisation, where a valid token continues to open a path that the organisation no longer intends to allow. If an attacker or insider gets hold of that token before expiry, they can often keep using it until the token naturally times out, even after access has been removed. That turns permission changes into a delayed-control problem instead of an immediate containment action.
Failure mechanism: The server trusts a self-contained token whose claims were true when issued, but cannot invalidate that token quickly when rights change, so downstream services continue to accept outdated authority.
Impact: Revoked users, compromised sessions, or former administrators may still reach sensitive data or perform restricted actions, which delays containment and can widen the effect of an account or token compromise.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 and MITRE ATT&CK address the attack and risk surface, while NIST SP 800-53 Rev 5 and NIST SP 800-63 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-07 — Long-Lived Secrets | JWTs that cannot be revoked behave like long-lived bearer secrets after rights change. |
| Recommendation — Shorten token lifetimes and rotate or invalidate bearer material when access changes. | ||
| NIST SP 800-53 Rev 5 | IA-5 — Authenticator Management | Token lifecycle and revocation are part of authenticator management for access control. |
| AC-2 — Account Management | Permission changes must propagate so removed access stops authorizing requests. | |
| AC-6 — Least Privilege | Stale JWTs can preserve excess privilege after a role or permission reduction. | |
| Recommendation — Define token issuance, renewal, revocation, and expiration handling for all authenticated access. Synchronize account changes with access enforcement and remove stale entitlements promptly. Limit token scopes and privileges so old tokens expose the smallest possible blast radius. | ||
| NIST SP 800-63 | Digital Identity Guidelines | The issue concerns authenticators and session validity after access changes. |
| Recommendation — Adopt identity session rules that limit how long an old assertion can remain usable. | ||
| MITRE ATT&CK | T1528 — Steal Application Access Token | Bearer tokens are attractive because possession can outlive current permission state. |
| Recommendation — Monitor for token theft and abuse patterns that let stolen tokens survive permission changes. | ||
Practitioner Guidance
What to verify: Confirm whether every sensitive path relies only on JWT signature and expiry, or whether there is an additional control for revocation, introspection, session state, or entitlement re-checking before high-risk actions. If the answer is “expiry only”, treat revocation gaps as an architectural exposure, not an edge case.
Decision rule: If the token can authorise production data access or administrative actions, keep its lifetime short enough that stale privilege is tolerable, and use a stronger invalidation path for offboarding or emergency access removal. If that is not possible, redesign the flow so the server re-evaluates current authority before the most sensitive requests.
Common mistake: Treating token revocation as unnecessary because the token is signed. Signature proves authenticity, not that the authority it represents is still current.
Practitioner takeaway: The key design choice is whether your system can safely tolerate stale privilege for the token’s full lifetime, because if it cannot, revocation or a live authorisation check is part of the control, not an optional enhancement.
Related resources from NHI Mgmt Group
- What breaks when security automation cannot re-test controls after change?
- What happens when SaaS permissions are not revoked promptly after employees change roles or leave?
- What breaks when cloud identity access is not revoked quickly after a role change or departure?
- What breaks when cloud permissions can change service behaviour after deployment?
Deepen Your Knowledge
Free weekly newsletter
Subscribe to the NHI & AI Identity Journal
The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.
Bonus 33% off our NHI Course when you subscribe.
Reviewed and updated by the NHIMG editorial team on October 31, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org