Skip to main content
Press slash or control plus K to focus the search. Use the arrow keys to navigate results and press enter to open a threat.
Reconnecting to live updates…

Malicious code in @yeaft/webchat-agent (npm)

0
Critical
Published: 07/16/2026 (07/16/2026, 18:44:35 UTC)
Source: GCVE Database
Product: @yeaft/webchat-agent

Description

The @yeaft/webchat-agent npm package installs a user-level service that opens a WebSocket connection to a configurable server. This connection allows remote commands to spawn an interactive shell as the installing user, effectively providing full host code execution. The package also supports remote self-upgrade from a publisher-controlled registry and clones a GitHub repository on startup to install executable plugin hooks, enabling persistent remote code execution. These behaviors constitute a backdoor with a high risk of compromise for any system running the affected versions.

Affected software

npmghsa
@yeaft/webchat-agent
Affected versions
=1.0.171=1.0.199=1.0.325=1.0.376=1.0.189=1.0.283=1.0.252=1.0.238=1.0.243=1.0.282=1.0.216=1.0.184=1.0.352=1.0.329=1.0.361=1.0.281=1.0.217=1.0.188=1.0.312=1.0.350=1.0.253=1.0.355=1.0.300=1.0.206=1.0.257=1.0.226=1.0.337=1.0.309=1.0.273=1.0.311=1.0.269=1.0.289=1.0.258=1.0.222=1.0.374=1.0.224=1.0.351=1.0.336=1.0.335=1.0.242=1.0.307=1.0.172=1.0.228=1.0.219=1.0.321=1.0.249=1.0.211=1.0.371=1.0.368=1.0.244=1.0.262=1.0.344=1.0.180=1.0.237=1.0.334=1.0.323=1.0.250=1.0.315=1.0.298=1.0.354=1.0.227=1.0.332=1.0.274=1.0.182=1.0.378=1.0.245=1.0.369=1.0.192=1.0.194=1.0.220=1.0.338=1.0.296=1.0.232=1.0.367=1.0.202=1.0.310=1.0.324=1.0.341=1.0.308=1.0.255=1.0.364=1.0.198=1.0.204=1.0.247=1.0.277=1.0.290=1.0.203=1.0.263=1.0.201=1.0.254=1.0.288=1.0.278=1.0.349=1.0.183=1.0.235=1.0.175=1.0.260=1.0.365=1.0.285=1.0.239=1.0.295=1.0.353=1.0.331=1.0.193=1.0.384=1.0.185=1.0.342=1.0.240=1.0.304=1.0.246=1.0.287=1.0.292=1.0.173=1.0.264=1.0.359=1.0.195=1.0.316=1.0.377=1.0.234=1.0.267=1.0.210=1.0.270=1.0.314=1.0.363=1.0.181=1.0.221=1.0.293=1.0.233=1.0.214=1.0.373=1.0.333=1.0.200=1.0.356=1.0.190=1.0.207=1.0.284=1.0.317=1.0.294=1.0.272=1.0.186=1.0.271=1.0.348=1.0.191=1.0.357=1.0.213=1.0.347=1.0.265

Run on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.

AI-Powered Analysis

Machine-generated threat intelligence

AILast updated: 08/06/2026, 18:40:36 UTC

Technical Analysis

The @yeaft/webchat-agent package installs itself as a user-level service (systemd on Linux, launchd/PM2 on other platforms) and opens a WebSocket to a configurable server URL. Through this WebSocket, the server can send commands to spawn the user's login shell and send input directly to it, granting interactive shell access to whoever controls the server or can intercept the WebSocket. The package also supports remote self-upgrade by running npm install from a publisher-controlled registry and restarting itself, allowing the remote party to swap the agent's code at will. Additionally, on every startup, it clones a GitHub repository from the main branch into a user directory, sets executable permissions on hook scripts, and enables these plugins, allowing remote code execution via plugin hooks on subsequent boots. The service environment prepends user-local binary paths to PATH, facilitating execution of user-installed binaries. This combination of remote shell access, self-upgrade capability, and plugin installation constitutes a backdoor with full host code execution as the installing user.

Potential Impact

An attacker controlling the remote server or able to intercept the WebSocket connection can gain interactive shell access with the privileges of the installing user, enabling full control over the host system. The remote self-upgrade feature allows the attacker to replace the agent's code arbitrarily, and the automatic cloning and enabling of plugins from a GitHub repository provides persistent remote code execution. This creates a severe security risk, effectively a backdoor that compromises the entire host environment where the package is installed.

Mitigation Recommendations

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix or mitigation is available, users should avoid installing or running the @yeaft/webchat-agent package and remove it from affected systems if present. Network controls to block WebSocket connections to untrusted servers and monitoring for unexpected service installations may help reduce exposure.

Pro Console: star threats, build custom feeds, automate alerts via Slack, email & webhooks.Upgrade to Pro

Technical Details

Gcve Source
db.gcve.eu
Osv Id
MAL-2026-10723
Osv Schema Version
1.7.4
Aliases
[]
Ecosystems
["npm"]
Database Specific Severity
null
Cvss Version
null

Threat ID: 6a74cf8ebf8831d5391aea46

Added to database: 08/06/2026, 18:16:46 UTC

Last enriched: 08/06/2026, 18:40:36 UTC

Last updated: 08/06/2026, 18:40:36 UTC

Views: 2

Community Reviews

0 reviews

Crowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.

Sort by
Loading community insights…

Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.

Actions

PRO

Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.

Please log in to the Console to use AI analysis features.

Need more coverage?

Upgrade to Pro Console for AI refresh and higher limits.

For incident response and remediation, OffSeq services can help resolve threats faster.

Latest Threats

Breach by OffSeqOFFSEQFRIENDS — 25% OFF

Check if your credentials are on the dark web

Instant breach scanning across billions of leaked records. Free tier available.

Scan now
OffSeq TrainingCredly Certified

Lead Pen Test Professional

Technical5-day eLearningPECB Accredited
View courses