registry / @agentvox/host / 0.1.39
@agentvox/host@0.1.39
AgentVox host daemon — runs on machines you want to control hands-free via the AgentVox voice app.
AI Security Review
scanned 16d ago · by lpm-firewall-aiSource inspection found package-aligned proxy communication, request-triggered agent activity reporting, and a narrowly scoped postinstall permission fix. The flagged decompression path stores board database data rather than executing a payload.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
The daemon pairs and connects, or receives an agent activity request.
Impact
The inspected behavior does not establish credential theft, arbitrary payload execution, or unconsented broad system mutation.
Mechanism
Host metadata goes to the configured proxy; agent activity reads local session history and returns it to the requester. The postinstall helper only adjusts a dependency helper's executable bit.
AI rationale
Source inspection found package-aligned proxy communication, request-triggered agent activity reporting, and a narrowly scoped postinstall permission fix. The flagged decompression path stores board database data rather than executing a payload.
Network endpoints2
proxy.agentvox.botregistry.npmjs.org
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The postinstall hook runs a helper that only makes node-pty's spawn helper executable when needed.
- The flagged CLI code reads Claude session history for agent activity; the activity handler runs on an agent.activity request and sends its result to the requester.
- The CLI and Vox send host metadata to the configured proxy as part of their connection; the default proxy is the package's AgentVox service.
- The Vox gzip path writes decoded bytes into its board database replica; it does not execute them.
Evidence against
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkWebSocket
HighEntropyStringsUrlStrings
WildcardDependency
Source & flagged code
10 flaggedpackage.jsonView file
•scripts.postinstall = node ./scripts/fix-pty-helper.mjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkgdist/vox.jsView file
13import { join } from "node:path";
L14: import { spawn, spawnSync } from "node:child_process";
L15: var REGISTRY = "https://registry.npmjs.org";
L16: var CHECK_INTERVAL_MS = 6 * 60 * 60 * 1e3;
...
L22: return null;
L23: const j2 = await r.json();
L24: return typeof j2.version === "string" ? j2.version : null;
...
L45: var defaultReExec = () => {
L46: const label = process.env.XPC_SERVICE_NAME;
L47: if (process.platform === "darwin" && label && /^[\w.-]+$/.test(label) && typeof process.getuid === "function") {
L48: try {
L49: spawn("/bin/sh", ["-c", `sleep 5; /bin/launchctl kickstart gui/${process.getuid()}/${label}`], {
Critical
Remote Asset Decode Execute
Source fetches a remote non-code asset, decodes its contents, and dynamically executes the decoded payload.
dist/vox.jsView on unpkg · L13Findings
3 Critical6 High5 Medium5 Low
CriticalRemote Asset Decode Executedist/vox.js
CriticalTrigger Reachable Dangerous Capabilitydist/vox.js
CriticalPrevious Version Dangerous Deltadist/cli.js
HighInstall Time Lifecycle Scriptspackage.json
HighChild Processdist/vox.js
HighHost Fingerprint Exfiltrationdist/vox.js
HighCredential Redirect Persistencedist/cli.js
HighSandbox Evasion Gated Capabilitydist/vox.js
HighRuntime Package Installdist/vox.js
MediumNetwork
MediumEnvironment Vars
MediumShips Build Helperscripts/install-systemd.sh
Medium
Affected versions and remediation
This report applies to @agentvox/host@0.1.39.
See all versions of @agentvox/host for other recorded verdicts.
- Review the evidence and your use of @agentvox/host@0.1.39 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.