Prow

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 Critical, reaching code or artifact modification (promoted to Tier 1) │ $10,000 │ $11,000 Critical, contained to the CI system (Tier 3) │ $2,500 │ $3,125 High (Tier 3) │ $1,250 │ $1,563 Medium (Tier 3) │ $250 │ $313 Low (Tier 3) │ $100 │ $125 Submit a vulnerability CVE history 0 Last 12 months 0 Lifetime Access control 40 % Auth bypass 25 % Code execution 20 % Info disclosure 13 % Latest advisories Loading the feed Overview Prow is not widely deployed, and that is precisely the point: it is the CI system that builds, tests and merges Kubernetes itself, along with most of the kubernetes-sigs ecosystem, Istio, Knative and OpenShift's CI, and it runs at prow.k8s.io against every pull request to the largest open source project in infrastructure. Its footprint is one installation and its blast radius is every Kubernetes cluster on earth: an attacker who can make Prow merge a commit, or who can execute inside the build cluster and reach the release signing and image-push credentials, owns the Kubernetes release artifacts, a supply-chain position comparable to owning a Linux distribution's build farm. It is also unusually reachable for a researcher, because Prow's entire configuration is public in the repo by design, so the target can be studied in full before a single packet is sent. The CNCF pays it on the Kubernetes program at HackerOne as Tier 3 infrastructure, where Critical is $2,500, but promotes a finding that reaches code or artifact modification to Tier 1 and up to $10,000, the only asset in the program that jumps two tiers on outcome. sorona.xyz adds 1.25× on top of whatever the CNCF awards, up to $1,000 more, plus a complexity bonus of up to $1,000 for a well-chained finding. Response times 1 business day First response 10 business days Triage 10 business days Bounty In scope Two assets, both bounty-eligible up to critical severity: the kubernetes/test-infra repository, listed as "Test infrastructure for the Kubernetes project", and prow.k8s.io as a live URL asset carrying the instruction "Please limit automated scanning to 1qps". Prow prices as Tier 3 infrastructure, $100 to $2,500, promoted to Tier 1 and $10,000 where the compromise reaches code or artifact modification The webhook ingress hook and the plugin set behind it, including the trust model that decides whose comment Prow will act on The command handlers a contributor can reach: /test , /retest , /ok-to-test , /lgtm , /approve and the OWNERS resolution underneath them The job lifecycle, meaning plank , sinker , crier , horologium and the pod utilities ( clonerefs , initupload , entrypoint , sidecar ) that check out and run untrusted pull-request code in the build cluster tide , which decides what actually merges into kubernetes/kubernetes , and deck , the web UI that serves job logs and artifacts at prow.k8s.io Configuration and credential handling: config.yaml and plugins.yaml loading, the ProwJob resource, GitHub token brokering through ghproxy , and the secrets a build cluster holds The rest of kubernetes/test-infra , since the whole repository is the asset: the CI configuration for the Kubernetes project along with the release and artifact-publishing tooling it carries Out of scope Broken link hijacking in documentation Community management tooling: mailing lists, Slack, calendars and similar Config and path disclosure without demonstrated credential leakage, which the program does not usually treat as a vulnerability Social engineering against maintainers Code in kubernetes-sigs/prow , since the paid asset for this project is kubernetes/test-infra Automated scanning of prow.k8s.io above the stated 1 qps limit, which the program asks you not to run Reports from a maintainer, or from the author or reviewer of the vulnerable code, which the program cannot award Most-valued findings Anything ending in code or artifact modification, whether by making tide merge a commit or by reaching release signing and image-push credentials from the build cluster, since that is the finding promoted to Tier 1 and $10,000 A trust-model bypass in hook that lets an untrusted contributor's comment fire /test or /ok-to-test , turning a pull request into execution inside the build cluster Approval bypass, meaning /lgtm or /approve honoured from an account OWNERS does not authorise Escape from an untrusted pull-request job into other jobs, service accounts or secrets in the build cluster GitHub token or secret leakage through job logs, crier notifications or the public configuration 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. The $10,000 ceiling applies only when the finding reaches code or artifact modification, which promotes it from Tier 3 to Tier 1; a Prow finding that stops at the CI system caps at $2,500 Automated scanning against prow.k8s.io is limited to 1 qps, which the program states in the asset instruction Config and path disclosures are typically not treated as vulnerabilities on their own without demonstrated credential leakage Both the source repo kubernetes/test-infra and the live deployment prow.k8s.io are paid assets, so a finding can be demonstrated against either The public disclosure date is negotiated with the Kubernetes Security Response Committee rather than set by you Program stats 141 Reports resolved $112k Bounties paid 60% Response efficiency Disclosure Coordinated through the Kubernetes Security Response Committee, which negotiates a public disclosure date with the reporter and prefers full disclosure once user mitigation is available. Findings against build infrastructure normally need credential rotation before anything is published, so expect a longer hold than a library bug; sorona.xyz keeps your finding private until that agreed date. 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