GHSA-xrqc-7xgx-c9vh RCE via ZipSlip and symbolic links in argoproj/argo-workflows

• Go github.com/argoproj/argo-workflows/v3: >= 3.7.0, < 3.7.5 → fixed in 3.7.5 • Go github.com/argoproj/argo-workflows/v3: < 3.6.14 → fixed in 3.6.14 • Go github.com/argoproj/argo-workflows: <= 2.5.3-rc4 (no fix yet)

Summary The patch deployed against CVE-2025-62156 is ineffective against malicious archives containing symbolic links. Details The untar code that handles symbolic links in archives is unsafe. Concretely, the computation of the link's target and the subsequent check are flawed: https://github.com/argoproj/argo-workflows/blob/5291e0b01f94ba864f96f795bb500f2cfc5ad799/workflow/executor/executor.go#L1034-L1037 PoC 1. Create a malicious archive containing two files: a symbolik link with path "./work/foo" and target "/etc", and a normal text file with path "./work/foo/hostname". 2. Deploy a workflow like the one in https://github.com/argoproj/argo-workflows/security/advisories/GHSA-p84v-gxvw-73pf with the malicious archive mounted at /work/tmp. 3. Submit the workflow and wait for its execution. 4. Connect to the corresponding pod and observe that the file "/etc/hostname" was altered by the untar operation performed on the malicious archive. The attacker can hence alter arbitrary files in this way. Impact The attacker can overwrite the file /var/run/argo/argoexec with a script of their choice, which will be executed at the pod's start.