{"id":"CVE-2026-22551","aliases":["EUVD-2026-37900","GHSA-qwjm-9c66-w4q4","UBUNTU-CVE-2026-22551"],"status":"published","title":"[Eclipse Theia] Data Exfiltration via Markdown Image Rendering in AI Chat","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","severity":{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-12","score":0.00312,"percentile":0.24428,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-201","CWE-829"],"published":"2026-06-18T14:32:01Z","modified":"2026-10-28T02:30:33Z","fixes":[{"ecosystem":"npm","package":"@theia/ai-chat-ui","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-chat","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-claude-code","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-code-completion","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-core","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-editor","fixed_in":["1.71.0"],"sources":["ghsa","osv"]},{"ecosystem":"npm","package":"@theia/ai-ide","fixed_in":["1.71.0"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-chat-ui","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-chat","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-claude-code","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-code-completion","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-core","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-editor","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"ghsa","ecosystem":"npm","package":"@theia/ai-ide","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-chat-ui","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-chat","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-claude-code","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-code-completion","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-core","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-editor","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"npm","package":"@theia/ai-ide","ranges":[{"type":"SEMVER","events":[{"introduced":"0"},{"fixed":"1.71.0"}]}]},{"source":"osv","ecosystem":"Ubuntu:16.04:LTS","package":"eclipse","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"}]}],"versions":["3.8.1-8"]},{"source":"osv","ecosystem":"Ubuntu:18.04:LTS","package":"eclipse","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"}]}],"versions":["3.8.1-10","3.8.1-11"]}],"products":[{"vendor":"eclipse","product":"theia"},{"vendor":"eclipse foundation","product":"eclipse theia"},{"vendor":"the eclipse foundation","product":"eclipse theia"}],"references":[{"url":"https://gitlab.eclipse.org/security/cve-assignment/-/work_items/115","tags":["Patch","Vendor Advisory","web"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-22551","tags":["advisory"]},{"url":"https://github.com/eclipse-theia/theia/issues/16892","tags":["web"]},{"url":"https://github.com/eclipse-theia/theia/pull/17364","tags":["web"]},{"url":"https://github.com/eclipse-theia/theia/commit/e3fdfe6992389bc5fa611058d00c39d7408508ed","tags":["web"]},{"url":"https://github.com/advisories/GHSA-qwjm-9c66-w4q4"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/22xxx/CVE-2026-22551.json","tags":["advisory"]},{"url":"https://github.com/eclipse-theia/theia","tags":["package"]},{"url":"https://ubuntu.com/security/CVE-2026-22551","tags":["report"]},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-22551","tags":["report"]}],"sources":{"cve":[{"id":"CVE-2026-22551","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T14:32:01.444Z","modified":"2026-06-18T15:25:13.612Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium","assessor":"cna"}],"cwes":["CWE-201","CWE-829"],"updated_at":"2026-11-02T13:39:05.585774Z"}],"euvd":[{"id":"EUVD-2026-37900","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T14:32:01Z","modified":"2026-06-18T15:25:13Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium"}],"updated_at":"2026-11-02T18:20:36.672794Z"}],"ghsa":[{"id":"GHSA-qwjm-9c66-w4q4","title":"[Eclipse Theia] Data Exfiltration via Markdown Image Rendering in AI Chat","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T18:35:22Z","modified":"2026-06-19T19:19:23Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium"}],"cwes":["CWE-201"],"updated_at":"2026-11-02T14:40:30.53004Z"}],"nvd":[{"id":"CVE-2026-22551","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T16:16:52.96Z","modified":"2026-10-26T11:10:00.157Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/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":"medium","assessor":"secondary"},{"version":"3.1","score":6.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N","level":"medium","assessor":"primary"}],"cwes":["CWE-201","CWE-829"],"updated_at":"2026-11-02T14:12:24.912854Z"}],"osv":[{"id":"CVE-2026-22551","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T14:32:01.444Z","modified":"2026-10-28T02:30:33.664655252Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium"}],"cwes":["CWE-201","CWE-829"],"updated_at":"2026-11-02T15:44:58.42038Z"},{"id":"GHSA-qwjm-9c66-w4q4","title":"[Eclipse Theia] Data Exfiltration via Markdown Image Rendering in AI Chat","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T18:35:22Z","modified":"2026-09-10T03:50:50.303567764Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium"}],"cwes":["CWE-201"],"updated_at":"2026-11-02T15:46:55.122811Z"},{"id":"UBUNTU-CVE-2026-22551","description":"In Eclipse Theia versions prior to 1.71.0, the AI chat rendered Markdown image tags from AI responses, triggering HTTP requests to arbitrary external URLs without restriction. Combined with prompt injection in a malicious workspace, an attacker could induce the AI agent to construct image URLs encoding sensitive information from the workspace or conversation context, exfiltrating it to attacker-controlled servers. The workspace trust enforcement introduced in v1.71.0 mitigates the documented attack chain by disabling AI features in untrusted workspaces.","status":"published","published":"2026-06-18T16:16:00Z","modified":"2026-06-24T11:01:29.013227652Z","cvss":[{"version":"4.0","score":6.7,"vector":"CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N","level":"medium"},{"version":"3.1","score":6.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:N/A:N","level":"medium"}],"updated_at":"2026-11-02T16:56:01.334185Z"}]}}