NewApp Monitoring is liveSign up free
AgentbeamBeta
Open-source package vulnerability · npm

CVE-2026-44688: [Eclipse Theia] Indirect Prompt Injection via Adversarial Workspace File and Directory Names in AI Chat

highCVSS 4.0 8.4GHSA-3jww-hxqj-wfq2

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).

Published Jun 18, 2026 · Updated Jun 19, 2026

Affected packages

@theia/ai-chat-ui

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-chat-ui

@theia/ai-chat

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-chat

@theia/ai-claude-code

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-claude-code

@theia/ai-code-completion

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-code-completion

@theia/ai-core

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-core

@theia/ai-editor

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-editor

@theia/ai-ide

Ecosystem
npm
Affected
< 1.71.0
Fixed in
1.71.0
Check
npm ls @theia/ai-ide

Advisory details

From the GitHub Advisory Database (CC BY 4.0).

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.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. No analysis has been written for this entry yet. Methodology

More npm vulnerabilities

All npm entries