# CVE Reservation Request <!-- There's help in the Eclipse Foundation Project Handbook https://www.eclipse.org/projects/handbook/#vulnerability-cve Note that this issue is configured (see the quick actions at the bottom) to be created as confidential. Note that a vulnerability does not need to actually be resolved before it is reported and that these reports can be revised as needed (reopen the issue to request changes). If you do not know how to fill certain fields, mark that in the comment and we will help you. You can delete the comments (or not). --> The Eclipse Foundation is a [Common Vulnerabilities and Exposures](https://cve.mitre.org/) (CVE) Numbering Authority. Creating this ticket initiates **reservation** of a CVE ID for the documented vulnerability. The reserved CVE ID will be posted in a comment below, and kept **confidential** until explicit publication request. > [!note] > To request CVE *publication*, please open a [CVE publication](https://gitlab.eclipse.org/security/cve-assignment/-/issues/new?issuable_template=CVE%20Publication%20Request) ticket. Please fill in the fields below to draft the CVE record. --- <!-- Required. Specify the project's name (e.g., "Eclipse Dash") and Eclipse Foundation ID, e.g., "technology.dash". --> ## CVE record information **Project name:** Eclipse Theia **Project id:** ecd.theia <!-- Required (for publication). Specify the version range as precisely as possible, e.g., "[3.0, 3.5.1]" or "[3.0, 3.5.1)". Note that using the standard range notion, square brackets are inclusive (i.e., that version is included in the range), and round brakets are exclusive (the vulnerability affects all versions up to but not including the named version). Multiple ranges can be provided. --> **Versions affected:** All versions of Eclipse Theia prior to 1.71.0 <!-- Required (for publication). The Common Weakness Enumeration (CWE) code comes from here: https://cwe.mitre.org/, e.g., "CWE-93: Improper Neutralization of CRLF Sequences ('CRLF Injection')". Multiple codes can be provided. --> **Common Weakness Enumeration (CWE):** - [CWE-1427: Improper Neutralization of Input Used for LLM Prompting](https://cwe.mitre.org/data/definitions/1427.html) - [CWE-829: Inclusion of Functionality from Untrusted Control Sphere](https://cwe.mitre.org/data/definitions/829.html) <!-- Optional. The Common Attack Pattern Enumerations and Classifications (CAPEC) code comes from here: https://capec.mitre.org/, e.g., "CAPEC-63: Cross-Site Scripting (XSS)". Multiple codes can be provided. --> **Common Attack Pattern Enumerations and Classifications (CAPEC):** <!-- Optional. Provide a Common Vulnerability Scoring System (CVSS). Note that if you do not provide this, then some agencies (e.g. NIST) will compute it on the project's behalf. Please be sure to include the CVSS version number, e.g., "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H". There's help here: https://www.first.org/cvss/calculator/4.0 --> **Common Vulnerability Scoring System:** [CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:A/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N](https://www.first.org/cvss/calculator/4.0) <!-- Required (for publication). The summary should start with the name of the project, e.g., "Eclipse Vert.x", then a description of the affected versions, followed by a description of the problem. The summary should be concise. For example, "In Eclipse Vert.x version 3.0 to 3.5.1, the HttpServer response headers and HttpClient request headers do not filter carriage return and line feed characters from the header value. This allow unfiltered values to inject a new header in the client request or server response." --> **Summary:** In Eclipse Theia versions prior to 1.71.0, the AI chat agent processed workspace file and directory names as part of its prompt context without distinguishing them from system instructions. An attacker could craft a malicious repository with adversarial directory or file names that, when analyzed by the AI agent, would cause the agent to follow attacker-controlled instructions (indirect prompt injection). Combined with other AI chat features available in untrusted workspaces, this enabled attack chains leading to data exfiltration via Markdown image rendering or arbitrary command execution via task definitions. <!-- Required (for publication). Include a link to the issue (e.g., GitHub Security Advisory) that's being used to track/resolve the issue. Other links that provide more information can be provided. For example, you may later publish the link to the fix commit. --> **Links:** - Fix: https://github.com/eclipse-theia/theia/pull/17364 - Tracking issue: https://github.com/eclipse-theia/theia/issues/16892 - Original report: https://gitlab.eclipse.org/security/vulnerability-reports/-/issues/331 <!-- Optional. Add the name or pseudonym of the person who has reported the issue. --> **Credits:** - Piotr Ryciak (https://gitlab.eclipse.org/void01) <!-- Quick actions will configure the state of the issue. Leave these. --> <!-- Keep this as the last line -->