In OpenClaw, the sandboxed image tool did not honor tools.fs.workspaceOnly=true for mounted paths resolved by the sandbox FS bridge. This allowed reading out-of-workspace mounted images (for example /agent/*) and forwarding those bytes to vision model providers.
Sandbox boundary bypass with confidentiality impact. In affected versions, read/write/edit respected workspace-only guardrails, but image could still load mounted out-of-workspace files and exfiltrate them via model requests.
openclaw (npm)<= 2026.2.22-2>= 2026.2.23 (released)2026.2.22-2workspaceOnly was enforced in sandbox file tools and apply_patch, but not propagated/enforced for image sandbox path resolution. The fix threads workspaceOnly into image-tool construction and asserts sandbox-root containment before loading media bytes.
dd9d9c1c609dcb4579f9e57bd7b5c879d0146b53OpenClaw thanks @tdjackey for reporting.
{
"cwe_ids": [
"CWE-200",
"CWE-284"
],
"github_reviewed": true,
"github_reviewed_at": "2026-03-04T19:13:48Z",
"nvd_published_at": "2026-03-19T22:16:32Z",
"severity": "MODERATE"
}