Karapace

Skip to content What this pays sorona.xyz pays 1.25× the program's reward, up to +$1,000 per finding, plus a complexity bonus of up to $1,000 for especially complex or well-chained bugs. Finding │ Program pays │ Through sorona.xyz P1: Critical (Aiven Open Source Repositories target group) │ $1,000 - $1,500 │ $1,250 - $1,875 P2: High │ $500 - $1,000 │ $625 - $1,250 P3: Medium │ $250 - $500 │ $313 - $625 P4: Low │ $50 - $250 │ $63 - $313 Submit a vulnerability CVE history 2 Last 12 months 2 Lifetime Access control 30 % Parsing/validation 25 % Auth bypass 25 % Info disclosure 15 % Latest advisories Loading the feed Overview Karapace is the Apache-2.0 schema registry and REST proxy for Apache Kafka, and the standard escape hatch for organisations that want Schema Registry and REST Proxy semantics without Confluent's Community Licence. It is the default registry inside every Aiven for Apache Kafka deployment, which makes it one of the most widely operated non-Confluent registries in production, and it appears independently across self-hosted Kafka platforms and Kubernetes-based Kafka stacks. Architecturally it sits in the highest-value position in a Kafka estate: it terminates HTTP from application clients and it holds the schema contracts that every producer and consumer trusts, so a compatibility-check or authorization bug there is a route to poisoning data for every downstream consumer at once. Aiven Ltd pays for reports through its Bugcrowd managed program, under the Aiven Open Source Repositories target group: P1 pays $1,000 to $1,500, down to $50 at P4. The higher managed-service tiers on the same program page cover Aiven's SaaS, not this repository. sorona.xyz adds 1.25× on top of whatever Aiven awards, up to $1,000 more, plus a complexity bonus of up to $1,000 for a well-chained finding. In scope What the program declares: vulnerabilities in Aiven open source code, with github.com/Aiven-Open listed as a target. Karapace sits at github.com/Aiven-Open/karapace, so the whole repository is covered at P1 $1,000 to $1,500, and the project's own SECURITY.md sends reporters to the same program. The Schema Registry API in full: the subject, version, config and compatibility endpoints, and the authorization model in front of them The Kafka REST proxy, meaning the HTTP-to-Kafka bridge that produces and consumes on a caller's behalf, including its consumer-group handling Avro, JSON Schema and Protobuf schema parsing, and the compatibility checking that decides whether a new schema may replace an old one Authentication and ACL enforcement, since Karapace is frequently the only thing standing between an HTTP client and a Kafka cluster The _schemas compacted-topic backing store and the consistency handling that keeps registry state coherent across instances The published container images and the Python package as distributed Out of scope Anything other than the most recent release, which is the only version the program will grade Anything also present in an upstream repository that Karapace forked from, which belongs with that upstream project Application-level denial of service, excluded by the program's baseline for all targets Theoretical findings with no working, Karapace-specific proof of concept Missing security headers, TLS configuration choices, and SPF, DKIM or DMARC best-practice reports Software-version banner disclosure and descriptive error messages Issues previously reported to Aiven through any channel other than Bugcrowd, which the program refuses on its face Most-valued findings Authorization or ACL bypass that lets an HTTP caller read or write a subject, or produce to and consume from a topic, that Karapace was configured to deny. That is the largest bug class here at 30% and the shape that reaches P1. Authentication bypass on the REST proxy, which turns Karapace into an unauthenticated path into the Kafka cluster behind it A compatibility-check flaw that admits a schema evolution the checker should have rejected, because every producer and consumer downstream trusts that decision and breaks on the data it lets through Memory or parser defects in the Avro, JSON Schema or Protobuf decoding of attacker-supplied schemas, which is untrusted structured input reaching a parser by design Cross-tenant leakage through the _schemas backing store or the registry cache, where one tenant's subjects or schema bodies become visible to another Program rules Submit through sorona.xyz first. We validate, coordinate with the program's security team, and add the payout uplift on top of whatever it awards. Only the most recent release counts: findings against older versions are not eligible A finding also present in an upstream repository Karapace forked from is not eligible and belongs upstream Application-level denial of service is out of scope for every target on this program Automated submission is barred; there must be a human in the loop on every report Bring a working, Karapace-specific proof of concept: theoretical findings are closed The program acknowledged a volume-related reward backlog on 21 July 2026, so expect payment to lag validation Disclosure Coordinated through Bugcrowd; the project's own SECURITY.md points straight at the same program and promises a response within two working days. Because Aiven will not accept an issue that reached it through any other channel first, file with Aiven through Bugcrowd directly and hold the detail until a fixed release is out. Safe harbor Sorona will not pursue legal action against you for good-faith research submitted to us in line with this policy. We cannot speak for the upstream project, its vendor, or any third party, and we do not grant authorization on their behalf. You are responsible for understanding and following the laws that apply to you, wherever you are, and for staying within any scope the project has defined. Submit a vulnerability