registry / @agentvault/agentvault / 0.23.56
@agentvault/agentvault@0.23.56
End-to-end encrypted messaging for AI agents: connects an agent to its owner, rooms and other agents on AgentVault.
AI Security Review
scanned 8d ago · by lpm-firewall-aiThe source resolves the credential-transfer hint to intended API authentication for a user-invoked publishing operation, with the destination configured by the caller and defaulting to the package’s service. The manifest’s only lifecycle hook is prepublishOnly, and the reviewed runtime and dependency hints do not establish a concrete attack.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
A user runs the CLI skill-publishing command or invokes the package’s runtime API.
Impact
The reviewed source does not establish unrelated credential disclosure, destructive behavior, or install-time mutation.
Mechanism
The CLI requires an API key and posts skill data to its configured API URL, which defaults to the AgentVault API host.
AI rationale
The source resolves the credential-transfer hint to intended API authentication for a user-invoked publishing operation, with the destination configured by the caller and defaulting to the package’s service. The manifest’s only lifecycle hook is prepublishOnly, and the reviewed runtime and dependency hints do not establish a concrete attack.
Network endpoints1
api.agentvault.chat
Decision evidence
public snapshotAI called this Clean at 92.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- package.json exposes the runtime entrypoints and has only a prepublishOnly lifecycle command, which runs the build.
- The CLI requires an API key for skill publishing, sends it with the request to its configured API URL, and defaults that URL to the AgentVault API host.
- Runtime defaults and the OpenClaw integration use the AgentVault API host; the integration also supports account configuration.
- The inspected dependency excerpts show WebSocket exports and Origin handling, MCP SDK metadata, and native notification code with option filtering.
Evidence against
- The inspected credential transfer is part of skill publishing to the configured API endpoint, rather than a separate fixed receiver.
- The scanner’s subprocess and network indicators correspond to CLI and service functions; the inspected source does not establish an unrelated credential transfer or install-time attack.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystem
Source & flagged code
10 flaggeddist/cli.jsView file
129Source sends the broad process environment to a literal external destination.
L129: // ../../node_modules/libsodium-sumo/dist/modules-sumo-esm/libsodium-sumo.mjs
L130: var __filename, __dirname, url, path, Module, Module, root, window_, crypto_, randomValuesStandard, crypto2, randomValueNodeJS, _Module, libsodium_sumo_default;
L131: var init_libsodium_sumo = __esm({
...
L224: Module2.useBackupModule = function() {
L225: return new Promise(function(resolve6, reject2) {
L226: var Module3 = {};
...
L277: quit_2 = (status, toThrow) => {
L278: process.exitCode = status;
L279: throw toThrow;
...
L291: xhr.responseType = "arraybuffer";
L292: xhr.send(null);
L293: return new Uint8Array(xhr.response);
Critical
Hardcoded Runtime Data Exfiltration
Source sends credentials or rich application records to a package-controlled external receiver enabled by default.
dist/cli.jsView on unpkg · L12993537});
L93538: import { execSync } from "node:child_process";
L93539: import { existsSync as existsSync2, readFileSync as readFileSync3, writeFileSync, chmodSync, readdirSync, statSync } from "node:fs";
High
93650try {
L93651: execSync("curl -sf --max-time 3 http://127.0.0.1:18789/", { stdio: "pipe" });
L93652: return { name: "Gateway port (18789)", status: "pass", message: "Responding" };
...
L93664: }
L93665: const home = process.env.HOME ?? "";
L93666: const warnings = [];
Findings
1 Critical7 High6 Medium5 Low
CriticalHardcoded Runtime Data Exfiltrationdist/cli.js
HighChild Processdist/cli.js
HighShell
HighSame File Env Network Executiondist/cli.js
HighCredential Redirect Persistencedist/index.js
HighCommand Output Exfiltrationdist/cli.js
HighCross File Remote Execution Contextdist/cli.js
HighSemantic Analysis Limiteddist/cli.js
MediumDynamic Requiredist/openclaw-entry.js
MediumNetwork
MediumEnvironment Vars
MediumOversized Source Filedist/index.js
MediumOversized Cli Entrypointdist/cli.js
Affected versions and remediation
This report applies to @agentvault/agentvault@0.23.56.
See all versions of @agentvault/agentvault for other recorded verdicts.
- Review the evidence and your use of @agentvault/agentvault@0.23.56 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.