{"id":"CVE-2026-26320","aliases":["EUVD-2026-8431","GHSA-7q2j-c4q5-rm27"],"status":"published","title":"OpenClaw macOS deep link confirmation truncation can conceal executed agent message","description":"OpenClaw is a personal AI assistant. OpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run.\" At the time of writing, the OpenClaw macOS desktop client is still in beta. In versions 2026.2.6 through 2026.2.13, an attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed. If a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message. The issue is fixed in 2026.2.14. Other mitigations include not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites and usingunattended deep links only with a valid `key` for trusted personal automations.","severity":{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-12","score":0.00492,"percentile":0.4275,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-451"],"published":"2026-02-17T21:41:40Z","modified":"2026-10-28T02:30:55Z","fixes":[{"ecosystem":"npm","package":"openclaw","fixed_in":["2026.2.14"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"npm","package":"openclaw","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"2026.2.6-0"},{"fixed":"2026.2.14"}]}]},{"source":"osv","ecosystem":"npm","package":"openclaw","ranges":[{"type":"SEMVER","events":[{"introduced":"2026.2.6-0"},{"fixed":"2026.2.14"}]}]}],"products":[{"vendor":"openclaw","product":"openclaw"}],"references":[{"url":"https://github.com/openclaw/openclaw/security/advisories/GHSA-7q2j-c4q5-rm27","tags":["x_refsource_CONFIRM","Mitigation","Patch","Vendor Advisory","advisory","web"]},{"url":"https://github.com/openclaw/openclaw/commit/28d9dd7a772501ccc3f71457b4adfee79084fe6f","tags":["x_refsource_MISC","Patch","fix","web"]},{"url":"https://github.com/openclaw/openclaw/releases/tag/v2026.2.14","tags":["x_refsource_MISC","Product","Release Notes","web"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-26320","tags":["advisory"]},{"url":"https://github.com/advisories/GHSA-7q2j-c4q5-rm27"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/26xxx/CVE-2026-26320.json","tags":["advisory"]},{"url":"https://github.com/openclaw/openclaw","tags":["package"]}],"sources":{"cve":[{"id":"CVE-2026-26320","title":"OpenClaw macOS deep link confirmation truncation can conceal executed agent message","description":"OpenClaw is a personal AI assistant. OpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run.\" At the time of writing, the OpenClaw macOS desktop client is still in beta. In versions 2026.2.6 through 2026.2.13, an attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed. If a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message. The issue is fixed in 2026.2.14. Other mitigations include not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites and usingunattended deep links only with a valid `key` for trusted personal automations.","status":"published","published":"2026-02-19T22:24:33.409Z","modified":"2026-02-20T15:41:16.051Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high","assessor":"cna"}],"cwes":["CWE-451"],"updated_at":"2026-11-02T13:38:20.971157Z"}],"euvd":[{"id":"EUVD-2026-8431","description":"OpenClaw is a personal AI assistant. OpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run.\" At the time of writing, the OpenClaw macOS desktop client is still in beta. In versions 2026.2.6 through 2026.2.13, an attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed. If a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message. The issue is fixed in 2026.2.14. Other mitigations include not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites and usingunattended deep links only with a valid `key` for trusted personal automations.","status":"published","published":"2026-02-19T22:24:33Z","modified":"2026-02-20T15:41:16Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"updated_at":"2026-11-02T19:48:37.056368Z"}],"ghsa":[{"id":"GHSA-7q2j-c4q5-rm27","title":"OpenClaw macOS deep link confirmation truncation can conceal executed agent message","description":"### Summary\nOpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run\".\n\nAt the time of writing, the OpenClaw macOS desktop client is still in beta.\n\nAn attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed.\n\n### Impact\nIf a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message.\n\n## Affected Versions\n- OpenClaw macOS desktop client versions \u003e= 2026.2.6 and \u003c= 2026.2.13.\n\n## Fixed Versions\n- 2026.2.14.\n\n### Mitigations\n- Do not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites.\n- Use unattended deep links only with a valid `key` for trusted personal automations.\n\n### Resolution\nUnkeyed deep links now enforce a strict message length limit for confirmation and ignore delivery/routing knobs (`deliver`, `to`, `channel`) unless a valid unattended `key` is provided.\n\nFix commit: 28d9dd7a772501ccc3f71457b4adfee79084fe6f\n\n---\n\nFix commit 28d9dd7a772501ccc3f71457b4adfee79084fe6f confirmed on main and in v2026.2.14. Upgrade to `openclaw \u003e= 2026.2.14`.","status":"published","published":"2026-02-17T21:41:40Z","modified":"2026-02-20T16:44:26Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-451"],"updated_at":"2026-11-02T14:39:25.82905Z"}],"nvd":[{"id":"CVE-2026-26320","description":"OpenClaw is a personal AI assistant. OpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run.\" At the time of writing, the OpenClaw macOS desktop client is still in beta. In versions 2026.2.6 through 2026.2.13, an attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed. If a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message. The issue is fixed in 2026.2.14. Other mitigations include not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites and usingunattended deep links only with a valid `key` for trusted personal automations.","status":"published","published":"2026-02-19T23:16:25.017Z","modified":"2026-06-17T10:26:05.203Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X","level":"high","assessor":"secondary"},{"version":"3.1","score":6.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:N","level":"medium","assessor":"primary"}],"cwes":["CWE-451"],"updated_at":"2026-11-02T14:12:06.274729Z"}],"osv":[{"id":"CVE-2026-26320","title":"OpenClaw macOS deep link confirmation truncation can conceal executed agent message","description":"OpenClaw is a personal AI assistant. OpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run.\" At the time of writing, the OpenClaw macOS desktop client is still in beta. In versions 2026.2.6 through 2026.2.13, an attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed. If a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message. The issue is fixed in 2026.2.14. Other mitigations include not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites and usingunattended deep links only with a valid `key` for trusted personal automations.","status":"published","published":"2026-02-19T22:24:33.409Z","modified":"2026-10-28T02:30:55.031864553Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-451"],"updated_at":"2026-11-02T15:44:59.80315Z"},{"id":"GHSA-7q2j-c4q5-rm27","title":"OpenClaw macOS deep link confirmation truncation can conceal executed agent message","description":"### Summary\nOpenClaw macOS desktop client registers the `openclaw://` URL scheme. For `openclaw://agent` deep links without an unattended `key`, the app shows a confirmation dialog that previously displayed only the first 240 characters of the message, but executed the full message after the user clicked \"Run\".\n\nAt the time of writing, the OpenClaw macOS desktop client is still in beta.\n\nAn attacker could pad the message with whitespace to push a malicious payload outside the visible preview, increasing the chance a user approves a different message than the one that is actually executed.\n\n### Impact\nIf a user runs the deep link, the agent may perform actions that can lead to arbitrary command execution depending on the user's configured tool approvals/allowlists. This is a social-engineering mediated vulnerability: the confirmation prompt could be made to misrepresent the executed message.\n\n## Affected Versions\n- OpenClaw macOS desktop client versions \u003e= 2026.2.6 and \u003c= 2026.2.13.\n\n## Fixed Versions\n- 2026.2.14.\n\n### Mitigations\n- Do not approve unexpected \"Run OpenClaw agent?\" prompts triggered while browsing untrusted sites.\n- Use unattended deep links only with a valid `key` for trusted personal automations.\n\n### Resolution\nUnkeyed deep links now enforce a strict message length limit for confirmation and ignore delivery/routing knobs (`deliver`, `to`, `channel`) unless a valid unattended `key` is provided.\n\nFix commit: 28d9dd7a772501ccc3f71457b4adfee79084fe6f\n\n---\n\nFix commit 28d9dd7a772501ccc3f71457b4adfee79084fe6f confirmed on main and in v2026.2.14. Upgrade to `openclaw \u003e= 2026.2.14`.","status":"published","published":"2026-02-17T21:41:40Z","modified":"2026-02-20T17:11:20.04002Z","cvss":[{"version":"4.0","score":7.1,"vector":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:N/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-451"],"updated_at":"2026-11-02T15:46:41.349637Z"}]}}