CVE-2026-26958: 
Grafana vulnerability analysis and mitigation

CVE-2026-26958 is an improper initialization vulnerability in the filippo.io/edwards25519 Go library affecting the (*Point).MultiScalarMult method. The flaw causes the method to fail to initialize its receiver point before performing elliptic curve scalar multiplication, producing incorrect or undefined results. All versions of filippo.io/edwards25519 prior to 1.1.1 are affected. It was independently reported by researchers WeebDataHoarder and shaharcohen1, disclosed by the maintainer on February 17, 2026, and published to the GitHub Advisory Database on February 18, 2026. The CVSS v4 base score is 1.7 (Low) (GitHub Advisory).

The root cause is CWE-665 (Improper Initialization): the (*Point).MultiScalarMult function in extra.go did not call v.Set(NewIdentityPoint()) to initialize the receiver before beginning its computation loop. If the receiver was an already-initialized point other than the identity, the result of the multi-scalar multiplication would be mathematically incorrect. More critically, if the receiver was a zero-value (uninitialized) Go struct, the behavior was undefined — and in practice, the method returned an invalid elliptic curve point that compared Equal to every other point, effectively breaking any equality check that relied on the result. The fix, applied in commit d1c650a, inserts a single line to set the receiver to the identity point before the computation begins (GitHub Commit, GitHub Advisory).

The primary impact is on the integrity and availability of cryptographic operations that use the MultiScalarMult API. An application calling this method on a non-identity or uninitialized receiver will receive a silently incorrect result, which could undermine signature verification, key agreement, or other cryptographic protocols built on Edwards25519. The most severe scenario is that the returned invalid point compares Equal to every valid point, which could allow an attacker to bypass point-equality checks in authentication or verification logic if they can influence the receiver state. Confidentiality impact is assessed as none, and the vulnerability does not directly enable remote code execution (GitHub Advisory).

No public exploit code or in-the-wild exploitation has been reported for this vulnerability. The EPSS score is approximately 0.018% (5th percentile), indicating a very low probability of exploitation in the near term. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. Exploitation is limited to applications that directly call the MultiScalarMult advanced API with a non-identity or uninitialized receiver — indirect dependencies (e.g., via github.com/go-sql-driver/mysql) are not affected (GitHub Advisory).

The maintainer released filippo.io/edwards25519 version 1.1.1 on February 17, 2026, which resolves the issue by initializing the receiver to the identity point at the start of MultiScalarMult. Developers should update their Go module dependency to filippo.io/edwards25519 v1.1.1 or later. Applications that do not directly call MultiScalarMult (e.g., those depending on the library only transitively through github.com/go-sql-driver/mysql) are not affected and do not require urgent action. The tool govulncheck is recommended for precise impact assessment (GitHub Release, GitHub Advisory).

The vulnerability was independently discovered and reported by two researchers, WeebDataHoarder and shaharcohen1, demonstrating community engagement with the library's security. The advisory explicitly notes that many downstream users (e.g., those depending on the library via MySQL drivers) are not affected, and recommends using govulncheck over less precise scanners to avoid alert fatigue — a notable commentary on the state of Go vulnerability scanning tooling. Downstream Linux distributions including SUSE and openSUSE issued security updates incorporating the fix (GitHub Advisory, SUSE Advisory).