Your RSA-2048 keys break in 2030. Find every one of them before attackers do.
📦
📦 npm
Not in CISA KEV
HIGH severity

GHSA-9h47-pqcx-hjr4

HIGH

GHSA-9h47-pqcx-hjr4 is a high-severity (CVSS 8.7) Broken Cryptographic Algorithm vulnerability in better-auth. O3 Security confirms whether GHSA-9h47-pqcx-hjr4 is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.

Better Auth has insecure cryptographic defaults in oidcProvider: alg=none advertised and plain PKCE accepted by default

Also known asCVE-2026-67336
Published
Jul 7, 2026
Updated
Aug 2, 2026
Affected
1 pkg
Patched
1 / 1
Exploits
None indexed
Exploitation data as of Sep 8, 2026 · OSV.dev, NVD, FIRST.org (EPSS)

Exploitation Status

No confirmed exploitation observed yet

  • CISA’s own triage has not observed active exploitation or public proof-of-concept code for this CVE as of its last assessment.

Exploitation and automatability from CISA’s SSVC triage for GHSA-9h47-pqcx-hjr4.

EPSS Exploitation Probability

via FIRST.org ↗
0.2%probability of exploitation in next 30 days
Lower Risk0.00%
Lower risk than most CVEs5th percentile — riskier than 5% of all scored CVEsHighest risk
0.00%0.22%0.44%0.66%0.2%0.2%Sep 26Sep 26

EPSS (Exploit Prediction Scoring System) is a daily probability model maintained by FIRST.org. It estimates the likelihood a CVE will be exploited in production environments within the next 30 days, derived from real-world threat intelligence signals.

How urgent is this, really

GHSA-9h47-pqcx-hjr4 plotted by exploitation likelihood (EPSS) against impact (CVSS). The shaded corner — EPSS 50%+ and CVSS 7.0+ — is where this CVE doesn't sit, though severity or exploitability alone can still warrant action.

Where this sits among everything scored

Of 369,023 CVEs with a current EPSS score, this one falls in the < 10% band (highlighted). Real counts from FIRST.org, not a sample — log-scaled since the landscape is heavily right-skewed.

Real-World Exposure

1 pkg affected

How broadly this vulnerability is actually deployed: weekly install volume shows current usage, and reverse-dependency count shows how many other packages break if it stays unpatched.

742other npm packages depend on this — each one inherits the vulnerability until it's patched upstream
better-authnpm
7.3Mdownloads / week

Description

Am I affected?

Users are affected if all of the following are true:

  • Their application uses better-auth at a version below the patched release.
  • Their application enables oidcProvider() from better-auth/plugins/oidc-provider or mcp() from better-auth/plugins/mcp (the mcp plugin delegates to oidcProvider and inherits both defaults).
  • For the algorithm-negotiation impact: relying parties of the application's OIDC server use a JWT verification library that performs algorithm negotiation from the discovery document without pinning to a specific signing algorithm.
  • For the PKCE impact: the authorization URL is exposed to any party other than the user agent and the application's OP.

If the application only uses @better-auth/oauth-provider (the canonical replacement) and have not enabled the legacy plugins, it is not affected. The new package's discovery document excludes none and its authorize schema rejects plain at parse time.

Fix:

  1. Upgrade to [email protected] or later.
  2. Migrate from the deprecated oidcProvider and mcp plugins to @better-auth/oauth-provider when feasible.
  3. If developers cannot upgrade their applications, see workarounds below.

Summary

The legacy oidcProvider and mcp plugins exhibit two related defects in their OIDC discovery and authorize surfaces.

The discovery document advertises "none" in id_token_signing_alg_values_supported (and, for mcp, in resource_signing_alg_values_supported on the OAuth protected-resource metadata). Any relying party that performs algorithm negotiation from this metadata without pinning to a real signing algorithm may accept unsigned tokens.

PKCE plain is enabled by default. The runtime gate in the authorize handler accepts code_challenge_method=plain under this default, and a missing code_challenge_method parameter is silently downgraded to "plain" before the allowlist check. Discovery advertises code_challenge_methods_supported: ["S256"], contradicting the runtime acceptance of plain. RFC 9700 §2.1.1 (OAuth 2.1) explicitly forbids plain.

Details

The metadata builders unconditionally inject "none" into the alg list. The runtime authorize gate is structured so a buggy client that strips the code_challenge_method parameter still enters the plain code path because the handler rewrites the missing value to "plain" before the allowlist check fires.

@better-auth/oauth-provider (the deprecation target for oidcProvider) is not affected by either defect. The metadata builder uses a JWSAlgorithms type union that structurally excludes "none". The authorize schema is code_challenge_method: z.literal("S256").optional(), which rejects plain at parse time.

Patches

Fixed in [email protected]. The legacy oidcProvider and mcp plugins now:

  • Drop "none" from id_token_signing_alg_values_supported (both plugins) and from resource_signing_alg_values_supported (mcp). Discovery no longer advertises the unsigned-token option.
  • Default allowPlainCodeChallengeMethod to false. A request that explicitly passes code_challenge_method=plain is rejected with invalid_request unless the integrator opts in.
  • Reject a code_challenge without an accompanying code_challenge_method instead of silently rewriting the missing value to plain. Clients that send code_challenge must also send code_challenge_method=S256.

Discovery and runtime behavior align on S256 only by default.

Integrators who must keep plain PKCE for legacy clients can restore the previous shape with oidcProvider({ allowPlainCodeChallengeMethod: true }) (and likewise for mcp). With the opt-in set, a request that omits code_challenge_method is treated as plain again, preserving backwards compatibility while keeping the secure default for everyone else. Both legacy plugins are deprecated long-term; the recommended migration is @better-auth/oauth-provider, which never advertised none or accepted plain PKCE.

Workarounds

If developers cannot upgrade their applications immediately:

  • Disable plain PKCE explicitly: set oidcProvider({ allowPlainCodeChallengeMethod: false }) (and the equivalent on mcp). Closes the runtime acceptance of plain even though the silent downgrade still rewrites missing methods.
  • Override the metadata to drop "none" from id_token_signing_alg_values_supported. For oidcProvider, pass metadata: { id_token_signing_alg_values_supported: ["RS256"] }. For mcp, set the same on options.oidcConfig.metadata. Verify by curling the .well-known endpoint.
  • Migrate to @better-auth/oauth-provider: the package is the deprecation target and is unaffected by both defects.

Impact

  • Algorithm-negotiation downgrade: relying parties that read the discovery document without pinning may accept unsigned (alg: "none") tokens.
  • Authorization-code interception: PKCE plain does not protect the authorization code if the URL leaks (Referer headers, browser history, screen capture, proxy logs). PKCE S256 is what protects against that exposure; with plain the protection is absent.
  • OAuth 2.1 / RFC 9700 non-conformance: deployments shipping the defaults are non-compliant with current standards.

Credit

Reported by @subhanUmer.

Resources

Affected Packages

1 total 1 fixed
EcosystemPackageVulnerable rangeFix
📦npmbetter-authall versions1.6.11

Detection & mitigation playbook

Open-source dependency

  1. Detect

    Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for better-auth. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.

  2. Fix

    Update better-auth to 1.6.11 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-9h47-pqcx-hjr4 is resolved across your whole dependency graph.

  3. Workarounds

    If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.

  4. How O3 protects you

    O3 pinpoints whether GHSA-9h47-pqcx-hjr4 is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.

Tailored to GHSA-9h47-pqcx-hjr4. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.

Frequently Asked Questions

### Am I affected? Users are affected if all of the following are true: - Their application uses `better-auth` at a version below the patched release. - Their application enables `oidcProvider()` from `better-auth/plugins/oidc-provider` or `mcp()` from `better-auth/plugins/mcp` (the mcp plugin delegates to `oidcProvider` and inherits both defaults). - For the algorithm-negotiation impact: relying parties of the application's OIDC server use a JWT verification library that performs algorithm negotiation from the discovery document without pinning to a specific signing algorithm. - For the PKC
O3 Security · Impact-Aware SCA

Is GHSA-9h47-pqcx-hjr4 in your dependencies?

O3 detects GHSA-9h47-pqcx-hjr4 across npm dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.

Scan my dependencies How O3 SCA works
GHSA-9h47-pqcx-hjr4: better-auth (High 8.7) | O3 Security