agentto
agentto connector — links your locally-installed AI agent CLIs through link.agentto.net to Agent to Dashboard (dash.agentto.net). Run `agentto` for a TUI.
This report shows the evidence for agentto@0.5.36. 1 of 2 known versions is blocked or flagged. See all versions.
OpenSSF/OSV Malicious Advisory
imported 3mo ago · by OpenSSF/OSVOpenSSF/OSV advisory MAL-2026-10725 confirms this npm version as malicious.
Advisory
MAL-2026-10725
Source
OpenSSF Malicious Packages via OSV
Summary
Malicious code in agentto (npm)
Details
agentto@0.5.36 opens a WebSocket to the hardcoded server wss://link.agentto.net and routes `terminal.*` RPC frames received from that server into a shell/PTY spawned on the installer's host. The `terminal.input` handler (host/terminal-service.mjs around line 135) base64-decodes `params.dataBase64` from remote messages and writes the bytes to a PTY spawned as `/bin/sh -l` (or the user's `$SHELL`); the RPC dispatcher forwards any `terminal.*` method arriving on the relay socket to this host terminal service. The host terminal is enabled by default and only disabled via the environment variable `AGENTTO_TERMINAL_ENABLED=0`. Any party controlling link.agentto.net therefore obtains interactive shell execution as the running user on every host that starts this connector with defaults.
References
Source & flagged code
0 flaggedNo flagged code excerpts are attached to this scan.
Findings
1 High
HighOsv Malicious Advisory
Affected versions and remediation
This report applies to agentto@0.5.36.
The versions list below shows the verdict for each recorded version.
- Avoid installing agentto@0.5.36. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Source advisory published: .
This report uses published external intelligence. The advisory does not provide a separate source-code analysis for each listed version.
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.