github.com/babylonlabs-io/babylon
Affected
- introduced
0, fixed1.1.0(semver)
Fixed versions
1.1.0
Fix
- Fixed version
1.1.0go get github.com/babylonlabs-io/[email protected]
Advisory
Also known as: GO-2025-3686
Babylon Finality Provider MsgCommitPubRandList replay attack
Summary
A high vulnerability exists in the Babylon protocol's x/finality module due to a lack of domain separation in signed messages, combined with insufficient validation in the MsgCommitPubRandList handler. Specifically, the handler does not enforce that the submitted Commitment field is 32 bytes long. This allows an attacker to replay a signature originally generated for a different message (e.g., a Proof-of-Possession in MsgCreateFinalityProvider) as a MsgCommitPubRandList. By crafting the message parameters, an attacker can use the typically 20-byte address bytes (from the PoP context) to form the StartHeight, NumPubRand, and a shorter-than-expected Commitment (e.g., 4 bytes). The replayed signature will pass verification for this crafted message, leading to the injection of an invalid PubRand commitment.
Impact
Successful exploitation of this vulnerability, specifically via the PoP signature replay, allows an attacker to store an invalid PubRand commitment (with a non-standard length, e.g., 4 bytes) for a targeted Finality Provider (FP). Despite the commitment itself being malformed, it's the associated StartHeight and NumPubRand (derived from the replayed address bytes and typically very large) that cause severe consequences
Future recommendations
To minimize future risk of such attacks, all finality providers should:
Finder
Vulnerability discovered by:
High Basis: The source's own band: HIGH
0, fixed 1.1.0 (semver)1.1.0
1.1.0go get github.com/babylonlabs-io/[email protected]Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: GitHub Advisory Database (reviewed advisories) · CC BY 4.0