Malicious code in zyr-agent (npm)
The zyr-agent npm package includes an AI coding agent that uses a default provider routing chat completions to a hardcoded, author-controlled endpoint. This endpoint can return commands that the agent executes automatically without user confirmation, including arbitrary shell commands via a registered bash tool. This creates a remote command execution channel controlled by the package author rather than the user. Two specific affected versions are identified: 1.5.7 and 1.6.2. No official patch or remediation guidance is provided yet.
AI Analysis
Technical Summary
zyr-agent ships an AI coding agent whose default 'free' provider routes chat completions to a hardcoded preview-slug endpoint controlled by the package author. The agent processes tool_calls from this endpoint and executes them automatically without user confirmation across CLI, Electron IPC, and HTTP server interfaces. One tool is a bash executor that runs commands synchronously on the host system. This design allows whoever controls the preview subdomain to execute arbitrary shell commands on any machine running the default provider. Additionally, a separate provider uses a hardcoded third-party OpenRouter API key, routing prompts through an author-controlled account. The combination of auto-execution without per-tool confirmation and a mutable, author-controlled endpoint creates a remote command execution vector controlled by the package author rather than the operator.
Potential Impact
An attacker controlling the hardcoded preview endpoint can execute arbitrary shell commands on any system running the affected zyr-agent versions with the default provider. This leads to full remote command execution capabilities without user interaction or confirmation, posing a critical security risk to affected hosts.
Mitigation Recommendations
No official patch or remediation guidance is currently available. Users should avoid running the affected versions (1.5.7 and 1.6.2) of zyr-agent or disable the default provider that routes to the author-controlled endpoint. Monitor vendor advisories for updates and patches addressing this issue.
Malicious code in zyr-agent (npm)
Description
The zyr-agent npm package includes an AI coding agent that uses a default provider routing chat completions to a hardcoded, author-controlled endpoint. This endpoint can return commands that the agent executes automatically without user confirmation, including arbitrary shell commands via a registered bash tool. This creates a remote command execution channel controlled by the package author rather than the user. Two specific affected versions are identified: 1.5.7 and 1.6.2. No official patch or remediation guidance is provided yet.
Affected software
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
Technical Analysis
zyr-agent ships an AI coding agent whose default 'free' provider routes chat completions to a hardcoded preview-slug endpoint controlled by the package author. The agent processes tool_calls from this endpoint and executes them automatically without user confirmation across CLI, Electron IPC, and HTTP server interfaces. One tool is a bash executor that runs commands synchronously on the host system. This design allows whoever controls the preview subdomain to execute arbitrary shell commands on any machine running the default provider. Additionally, a separate provider uses a hardcoded third-party OpenRouter API key, routing prompts through an author-controlled account. The combination of auto-execution without per-tool confirmation and a mutable, author-controlled endpoint creates a remote command execution vector controlled by the package author rather than the operator.
Potential Impact
An attacker controlling the hardcoded preview endpoint can execute arbitrary shell commands on any system running the affected zyr-agent versions with the default provider. This leads to full remote command execution capabilities without user interaction or confirmation, posing a critical security risk to affected hosts.
Mitigation Recommendations
No official patch or remediation guidance is currently available. Users should avoid running the affected versions (1.5.7 and 1.6.2) of zyr-agent or disable the default provider that routes to the author-controlled endpoint. Monitor vendor advisories for updates and patches addressing this issue.
Technical Details
- Gcve Source
- db.gcve.eu
- Osv Id
- MAL-2026-13471
- Osv Schema Version
- 1.7.4
- Aliases
- []
- Ecosystems
- ["npm"]
- Database Specific Severity
- null
- Cvss Version
- null
Threat ID: 6a7573b5bf8831d539d93adb
Added to database: 08/07/2026, 05:57:09 UTC
Last enriched: 08/07/2026, 07:42:46 UTC
Last updated: 08/07/2026, 07:42:46 UTC
Views: 2
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
External Links
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
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.