Answer in brief
CVE-2026-46580 records a High severity (CVSS 8.8) vulnerability in [Eclipse Theia] Indirect Prompt Injection via Auto-Loaded Workspace Prompt Template Files in AI Chat. The current sources do not mark it as known exploited. The current feed maps @theia/ai-chat (npm), @theia/ai-chat (npm), @theia/ai-chat-ui (npm), @theia/ai-chat-ui (npm) and additional mapped packages. Check affected ranges and fixed versions before updating.
Analysis pending evidence review
HOL Guard separates source facts from reviewed analysis. See the methodology.
CVSS is 8.8. The current sources do not mark it as known exploited. Treat this as a source-backed prioritization signal, not a statement about your environment.
Analysis status
Analysis pending evidence review
Factual feed record only; HOL analysis is not approved for indexing. Read the methodology.
The current feed maps @theia/ai-chat (npm), @theia/ai-chat (npm), @theia/ai-chat-ui (npm), @theia/ai-chat-ui (npm) and additional mapped packages. Check affected ranges and fixed versions before updating.
| Package | Affected range | Fixed version |
|---|---|---|
| @theia/ai-chatnpm | <1.71.0 | 1.71.0 |
| @theia/ai-chatnpm | >=0 <1.71.0 | 1.71.0 |
| @theia/ai-chat-uinpm | <1.71.0 | 1.71.0 |
| @theia/ai-chat-uinpm | >=0 <1.71.0 | 1.71.0 |
| @theia/ai-claude-codenpm | <1.71.0 | 1.71.0 |
| @theia/ai-claude-codenpm | >=0 <1.71.0 | 1.71.0 |
| @theia/ai-code-completionnpm | <1.71.0 | 1.71.0 |
| @theia/ai-code-completionnpm | >=0 <1.71.0 | 1.71.0 |
| @theia/ai-corenpm | <1.71.0 | 1.71.0 |
| @theia/ai-corenpm | >=0 <1.71.0 | 1.71.0 |
| @theia/ai-editornpm | <1.71.0 | 1.71.0 |
| @theia/ai-editornpm | >=0 <1.71.0 | 1.71.0 |
Published upstream
Jun 18, 2026
Evidence: source:osv:source_dates:source-dates:recordSource modified
Sep 10, 2026
Evidence: source:osv:source_dates:source-dates:recordFirst seen by HOL
Jun 19, 2026
In Eclipse Theia versions prior to 1.71.0, files matching the pattern .prompts/*.prompttemplate in a workspace were automatically loaded and could override or extend the AI agent's system prompts. An attacker could craft a malicious repository containing prompt template files that, when the workspace was opened in Theia, replaced the AI's system instructions with attacker-controlled content (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.
Quoted source text, attributed separately from HOL analysis.