A cryptographic key stored in a way that allows signing operations but prevents export of the raw private material. In browser-based DPoP, this is the practical control that keeps a stolen script from simply copying the proofing key and replaying access.
What the term means in practice
A non-extractable key is a key that can be used for cryptographic operations such as signing, but whose raw private material cannot be exported. That distinction matters because the protection is about preventing key theft, not just hiding the application logic that uses the key.
In browser-based proof-of-possession designs, the point is to keep the proofing key bound to the browser context so a stolen script or copied storage cannot simply lift the private key and replay it elsewhere. The security value comes from limiting what an attacker can reuse after access to the page or storage layer.
Why non-extractability matters for browser security
Non-extractable keys reduce the blast radius of token theft, XSS-driven exfiltration, and other browser-side compromise scenarios. If an attacker can read the key material directly, they can often move from one-time access to persistent replay or impersonation, especially when the same key is trusted for repeated proofs.
Browser Storage Security Guide explains the surrounding trade-offs in browser storage and why DPoP depends on keeping the proofing key non-extractable. The key point is that storage hardening and key non-exportability solve different parts of the same problem.
A browser implementation can still fail if the surrounding session, origin, or script environment is compromised. Non-extractability helps preserve the cryptographic boundary, but it does not prevent an attacker from using the key through the legitimate runtime while they control it.
Where this control fits in DPoP and similar proofs
In DPoP, the client proves possession of a private key when presenting a token or request proof. A non-extractable key strengthens that model because the proof can be generated without ever exposing the private key material to JavaScript or to easy copy-and-paste reuse.
This is most useful when the security goal is to make stolen bearer-style artefacts less reusable. The proof key becomes a binding factor, so an attacker who steals only the token cannot automatically present the same proof elsewhere unless they also control the runtime that can sign with the bound key.
The concept also generalises to any design that needs local signing without raw key export, such as client certificates, device-bound credentials, and WebCrypto-backed application keys. The common requirement is that the application can ask the key to sign, but cannot read the private bytes themselves.
Operational limits and lifecycle considerations
Non-extractable does not mean immutable, invisible, or automatically safe. The key still has a lifecycle, still depends on origin trust, and still needs revocation or replacement if the browser profile, session, or application trust boundary is lost.
NIST SP 800-57 Key Management is the right reference for thinking about key lifecycle, cryptoperiods, and replacement discipline, while the browser-side implementation determines whether the material is exportable in the first place. For the broader browser storage and logout pattern, Browser Storage Security Guide remains the most direct navigation aid.
Architecturally, the control is strongest when paired with short-lived proofs, origin isolation, and clear logout or re-authentication behaviour. That combination reduces the chance that a long-lived browser key survives past the moment it should have become useless.
Risk and Threat Considerations
Non-extractable keys primarily reduce key theft and replay risk, but they do not eliminate browser compromise. If an attacker gains script execution, DOM control, or browser runtime access, they may still use the key through the legitimate API surface even though they cannot export the raw private material.
Failure mechanism: the attacker abuses the trusted runtime to request signatures, turning a local non-exportable key into an on-box signing oracle while avoiding raw key extraction.
Impact: the attacker may sustain token replay, session abuse, or request forgery for as long as the bound key and session remain trusted, which can make post-compromise cleanup harder than simple token rotation alone.
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 addresses the attack and risk surface, while NIST SP 800-57 and NIST SP 800-53 Rev 5 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| NIST SP 800-57 | Key Management | Defines lifecycle and cryptoperiod handling for keys that must stay protected. |
| Recommendation — Set key lifetime, rotation, and destruction rules for browser-held signing keys. | ||
| NIST SP 800-53 Rev 5 | IA-5 — Authenticator Management | Covers issuance and lifecycle handling for authenticators and related credential material. |
| Recommendation — Manage browser-bound proof keys as controlled authenticators with defined lifecycle handling. | ||
| OWASP Non-Human Identity Top 10 | NHI-02 — Secret Leakage | Addresses prevention of private material exposure for non-human or application-held credentials. |
| NHI-07 — Long-Lived Secrets | Applies when a browser-held proof key persists longer than operationally safe. | |
| NHI-04 — Insecure Authentication | Relevant because the key is used to prove possession in a browser authentication flow. | |
| Recommendation — Prevent export paths that would expose private key material to scripts or storage theft. Keep browser-held signing keys short-lived and replace them promptly after risk events. Bind proof-of-possession flows to keys that cannot be exported or trivially replayed. | ||
Practitioner Guidance
What to watch for: treat non-extractability as one control in a larger browser trust model, not as a substitute for origin hardening, XSS prevention, or session hygiene. The useful question is whether the implementation still behaves safely when the page is partially compromised but the key cannot be copied out.
Governance implication: teams should define when a browser-held key is acceptable, how long it may live, and what event forces re-issuance or logout. That policy matters because a non-extractable key can still be operationally stale even when it remains technically protected.
Related resources from NHI Mgmt Group
- Who is accountable when an enclave key policy allows both attested and non-attested decrypt paths?
- Who is accountable for non-human identity risk when a service account or API key is over-permissioned?
- What breaks when teams rely on key rotation alone to secure non-human identities?
- What is the difference between encryption key management and access key management for non-human identities?
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 November 2, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org