Key Insights:
- The Solana Foundation has uncovered a security vulnerability within its token system that could permit unauthorized creation or withdrawal of tokens.
- The flaw is linked to the ZK ElGamal Proof program and impacts confidential transactions while leaving standard SPL tokens unaffected.
The Solana Foundation recently identified a critical vulnerability in its privacy-oriented token system, posing a risk of attackers creating fraudulent zero-knowledge proofs (ZKPs) to gain unauthorized access for minting or withdrawing tokens.
The issue was initially flagged on April 16 through Anza’s GitHub security advisory, including a functioning proof-of-concept. Rapid responses were initiated by engineers from Solana teams—Anza, Firedancer, and Jito—to remedy the breach, as outlined in a post-mortem published on May 2.
The defect was traced back to the ZK ElGamal Proof system which facilitates zero-knowledge proofs used in Solana’s Token-22 categorized confidential transfers, allowing for private transaction details by encryption while maintaining their validity via cryptographic methods.
Zero-knowledge proofs enable individuals to validate they possess certain resources—be it a password or an asset—without disclosing the specific item. In cryptocurrency contexts, they prove a transaction’s legitimacy while keeping amounts and addresses confidential, hence safeguarding against malicious planning.
The vulnerability arose from the omission of critical algebraic components during the Fiat-Shamir transformation—a recognized method to convert zero-knowledge proofs into a single verifiable proof.
A highly skilled attacker could forge false proofs that would be accepted by the on-chain verifier. This flaw could have led to the unauthorized minting of an infinite supply of tokens or withdrawal of tokens from alternate accounts, though it did not compromise the standard SPL tokens or the principal Token-2022 program logic.
From April 17 onwards, private patches were shared with validator operators, followed by an additional patch that same evening addressing another related issue. Both fixes underwent evaluations by independent security firms: Asymmetric Research, Neodyme, and OtterSec. As of April 18, a significant majority of validators had adopted the proposed solution.
As per the post-mortem’s findings, there remains no evidence that this bug was exploited; all assets are safe.
