You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Research correct CRE mapping for K09 (Misconfigured Cluster Components) in OWASP Kubernetes Top Ten 2022 #916
This issue tracks follow-up research to find the correct, distinct CRE identifiers for K09: Misconfigured Cluster Components in the OWASP Kubernetes Top Ten 2022 dataset.
Background
In PR #877, the owasp_kubernetes_top10_2022.json mapping for K09 was intentionally left unchanged because it currently shares the same cre_ids as K01 (["233-748", "486-813"]). After reviewing the upstream OWASP source, a clearly correct distinct replacement could not be identified, so a speculative change was avoided to prevent introducing incorrect data.
K09 is focused on hardening and misconfiguration of core cluster components (e.g., kubelet, etcd, kube-apiserver), with prevention guidance centred on secure configuration, CIS benchmark scans, and reducing unsafe defaults.
Open question
Is there a more specific CRE mapping for cluster-component hardening/misconfiguration that should replace or augment the current ["233-748", "486-813"] entries for K09?
Note that the OWASP Kubernetes Top Ten 2025 appears to consolidate this area under K07: Misconfigured And Vulnerable Cluster Components — the 2025 mapping (PR Add refresh scripts for OWASP resources for issue 471 #877, owasp_kubernetes_top10_2025.json) may provide useful pointers for the correct CRE IDs.
If no better mapping exists, the duplication with K01 should be explicitly documented in the data file (e.g., via a comment or a companion note) so future maintainers understand it is intentional.
#!/bin/bash
gh pr list --repo OWASP/OpenCRE --search "916" --state all
gh pr list --repo OWASP/OpenCRE --search "K09" --state all
gh pr list --repo OWASP/OpenCRE --search "kubernetes" --state all --limit 30
#!/bin/bashset -euo pipefail
body=$(cat <<'EOF'## ContextThis issue tracks follow-up research to find the correct, distinct CRE identifiers for **K09: Misconfigured Cluster Components** in the OWASP Kubernetes Top Ten 2022 dataset.## BackgroundIn PR `#877`, the `owasp_kubernetes_top10_2022.json` mapping for K09 was intentionally left unchanged because it currently shares the same `cre_ids` as K01 (`["233-748", "486-813"]`). After reviewing the upstream OWASP source, a clearly correct distinct replacement could not be identified, so a speculative change was avoided to prevent introducing incorrect data.**K09** is focused on hardening and misconfiguration of core cluster components (e.g., `kubelet`, `etcd`, `kube-apiserver`), with prevention guidance centred on secure configuration, CIS benchmark scans, and reducing unsafe defaults.## Open questionIs there a more specific CRE mapping for cluster-component hardening/misconfiguration that should replace or augment the current `["233-748", "486-813"]` entries for K09?## Guidance for review- Compare K09's prevention/mitigation content on the [OWASP Kubernetes Top Ten 2022 page](https://owasp.org/www-project-kubernetes-top-ten/2022/en/src/K09-misconfigured-cluster-components) against the CRE database.- Note that the OWASP Kubernetes Top Ten 2025 appears to consolidate this area under **K07: Misconfigured And Vulnerable Cluster Components** — the 2025 mapping (PR `#877`, `owasp_kubernetes_top10_2025.json`) may provide useful pointers for the correct CRE IDs.- If no better mapping exists, the duplication with K01 should be explicitly documented in the data file (e.g., via a comment or a companion note) so future maintainers understand it is intentional.## References- PR `#877`: https://github.com/OWASP/OpenCRE/pull/877- PR `#927` (implementation for this issue): https://github.com/OWASP/OpenCRE/pull/927- Comment thread: https://github.com/OWASP/OpenCRE/pull/877#discussion_r3351547169- OWASP K09 source: https://owasp.org/www-project-kubernetes-top-ten/2022/en/src/K09-misconfigured-cluster-components- OWASP K07 (2025 consolidation): https://owasp.org/www-project-kubernetes-top-ten/EOF)
gh issue edit 916 --repo OWASP/OpenCRE --body "$body"
gh issue view 916 --repo OWASP/OpenCRE
Length of output: 4920
@Bornunique911, I've updated the References section of issue #916 to add a link to PR #927 ("Research k09 mapping" by @SurbhiAgarwal1), which implements the fix (adding the "note" field to document the intentional K09/K01 CRE duplication) and closes this issue.
Note: since this thread is on an issue rather than a pull request, the @coderabbitai review command doesn't apply here — that command triggers a code review only on pull request diffs. If you'd like a review of the actual code changes, please post @coderabbitai review as a comment on PR #927 directly, and I'll review it there.
Activity
SurbhiAgarwal1 commentedon Jun 10, 2026
Hi @Pa04rth , Can you assign me this issue?
Pa04rth commentedon Jun 11, 2026
@SurbhiAgarwal1 Assigned !
Bornunique911 commentedon Jul 8, 2026
Please modify references accordingly with the new PR's.
@coderabbitai review
coderabbitai commentedon Jul 8, 2026
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash gh issue view 916 --repo OWASP/OpenCRE --commentsLength of output: 577
🏁 Script executed:
#!/bin/bash gh issue view 916 --repo OWASP/OpenCRELength of output: 2563
🏁 Script executed:
Length of output: 1986
🏁 Script executed:
#!/bin/bash gh pr view 927 --repo OWASP/OpenCRE --json title,body,url,filesLength of output: 2900
🏁 Script executed:
Length of output: 4920
@Bornunique911, I've updated the References section of issue#916to add a link to PR#927("Research k09 mapping" by@SurbhiAgarwal1), which implements the fix (adding the"note"field to document the intentional K09/K01 CRE duplication) and closes this issue.Note: since this thread is on an issue rather than a pull request, the
@coderabbitai reviewcommand doesn't apply here — that command triggers a code review only on pull request diffs. If you'd like a review of the actual code changes, please post@coderabbitai reviewas a comment on PR#927directly, and I'll review it there.