@agentvault/agentvault@0.23.21
The official **OpenClaw** plugin for AgentVault. Adds cryptographic identity, earned trust, and Signal-grade encrypted communications to OpenClaw agents. (Looking for a different framework? See **[Which Framework?](#which-framework)** below.)
AI Security Review
scanned 3mo ago · by lpm-firewall-aiNo malicious install-time action or exfiltration chain was confirmed. The package nevertheless performs explicit user-command OpenClaw provisioning, remote-managed workspace configuration, and persistence setup, warranting a warning.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
User runs agentvault setup, create, pm2-setup, or doctor --fix
Impact
User-authorized agent control-surface changes can affect the configured OpenClaw runtime.
Mechanism
explicit OpenClaw agent provisioning and remote configuration injection
AI rationale
No malicious install-time action or exfiltration chain was confirmed. The package nevertheless performs explicit user-command OpenClaw provisioning, remote-managed workspace configuration, and persistence setup, warranting a warning.
Evidence
package.jsonopenclaw.plugin.jsondist/cli.jsdist/openclaw-entry.js~/.openclaw/openclaw.json~/.openclaw/workspace-<name>/IDENTITY.md~/.openclaw/workspace-<name>/SOUL.md~/.openclaw/agentvault
Network endpoints3
api.agentvault.chat127.0.0.1:18789127.0.0.1:18790
Decision evidence
public snapshotAI called this Suspicious at 90.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- dist/cli.js explicit create/setup writes ~/.openclaw config and agent workspace files
- dist/cli.js create fetches remote hub configuration and writes IDENTITY.md/SOUL.md
- dist/cli.js setup can install the plugin, configure pm2 persistence, and restart OpenClaw
Evidence against
- package.json has no preinstall/install/postinstall lifecycle hook
- Mutating behavior is reached only by explicit CLI setup/create/doctor commands
- Network use targets configured AgentVault API or localhost; no unrelated receiver found
- Command execution is OpenClaw/pm2 diagnostics and requested setup, not command-output upload
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
Source & flagged code
12 flaggeddist/cli.jsView file
35Source sends the broad process environment to a literal external destination.
L35: // ../../node_modules/libsodium-sumo/dist/modules-sumo-esm/libsodium-sumo.mjs
L36: var __filename, __dirname, url2, path, Module, Module, root, window_, crypto_, randomValuesStandard, crypto4, randomValueNodeJS, _Module, libsodium_sumo_default;
L37: var init_libsodium_sumo = __esm({
...
L112: _Module = Module;
L113: Module.ready = new Promise(function(resolve4, reject) {
L114: var Module2 = _Module;
...
L183: quit_2 = (status, toThrow) => {
L184: process.exitCode = status;
L185: throw toThrow;
...
L197: xhr.responseType = "arraybuffer";
L198: xhr.send(null);
L199: 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 · L3535// ../../node_modules/libsodium-sumo/dist/modules-sumo-esm/libsodium-sumo.mjs
L36: var __filename, __dirname, url2, path, Module, Module, root, window_, crypto_, randomValuesStandard, crypto4, randomValueNodeJS, _Module, libsodium_sumo_default;
L37: var init_libsodium_sumo = __esm({
...
L112: _Module = Module;
L113: Module.ready = new Promise(function(resolve4, reject) {
L114: var Module2 = _Module;
...
L183: quit_2 = (status, toThrow) => {
L184: process.exitCode = status;
L185: throw toThrow;
...
L197: xhr.responseType = "arraybuffer";
L198: xhr.send(null);
L199: return new Uint8Array(xhr.response);
Critical
Command Output Exfiltration
Source executes local commands and sends command output to an external endpoint.
Findings
3 Critical6 High7 Medium4 Low
CriticalHardcoded Runtime Data Exfiltrationdist/cli.js
CriticalCommand Output Exfiltrationdist/cli.js
CriticalTrigger Reachable Dangerous Capabilitydist/cli.js
HighChild Processdist/cli.js
HighShell
HighSame File Env Network Executiondist/cli.js
HighCross File Remote Execution Contextdist/cli.js
HighTrigger Reachable Persistencedist/cli.js
HighSemantic Analysis Limiteddist/cli.js
MediumDynamic Requiredist/openclaw-entry.js
MediumNetwork
MediumEnvironment Vars
MediumInstall Persistencedist/cli.js
Affected versions and remediation
This report applies to @agentvault/agentvault@0.23.21.
See all versions of @agentvault/agentvault for other recorded verdicts.
- Review the evidence and your use of @agentvault/agentvault@0.23.21 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.