CloudPirates Helm Charts CVE-2026-45132

Exploitation required the CloudPirates/helm-charts repository to be on a commit prior to fcf9302 with the generate-schema.yaml (and pull-request.yaml) workflows configured to trigger on pull_request_target while checking out the PR head ref from forks and exposing the Personal Access Token plus SSH signing key as workflow secrets. Any unauthenticated GitHub user could open a pull request from a fork to trigger the workflow - no special privileges, user interaction by maintainers, or non-default GitHub configuration was needed beyond the repository being public and accepting external PRs. The limiting factor is that this is a single specific repository's CI environment; the vulnerability is not present in any deployed product, and post-patch the 'external' environment gate now requires reviewer approval before the privileged job runs.

The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:N) yields a 10.0 base score driven by network reach, no authentication, no user interaction, and a scope change reflecting that stolen credentials grant access beyond the original repo (the leaked PAT and SSH signing key can sign commits and push to other resources). Exploitation status is 'no public exploit identified at time of analysis' - neither KEV nor a published POC is referenced - though the underlying pull_request_target abuse pattern is widely documented and trivially automatable. EPSS data was not supplied, so probabilistic exploitation likelihood cannot be quantified. The realistic risk is high: any GitHub user can open a PR against a public repo, so the barrier to attempt exploitation is essentially zero, but exploitation is now historical because the maintainers have patched the workflow.

An attacker forks the CloudPirates/helm-charts repository, modifies the schema generation script or any file executed by the vulnerable workflow to exfiltrate environment secrets (e.g. curl the PAT and SSH signing key to an attacker-controlled endpoint), and opens a pull request against the upstream repo. The pull_request_target trigger fires automatically, GitHub Actions checks out the attacker's fork code, runs it with the privileged base-repo secrets in scope, and the attacker captures the credentials - which can then be used to push signed commits or access other resources the PAT authorizes.