registry / @cloop/runner / 0.3.37
@cloop/runner@0.3.37
Make your computer and your agents Cloop machines: device login, the machine daemon, the Cloop MCP server, and Claude Code hooks.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe manifest contains no install lifecycle hook, and the flagged command execution is reached through the user-selected `cloop connect` workflow. Its detected downloads install coding agents from their official vendor sources, consistent with the package’s stated purpose.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The installer and agent setup paths run when a user invokes the relevant CLI command, especially `cloop connect`.
Impact
Expected setup changes and agent installation; no unrelated data collection, credential theft, or hidden install-time execution was identified.
Mechanism
The connect workflow can install missing coding agents using their vendors’ installation paths and configure the Cloop machine.
AI rationale
The manifest contains no install lifecycle hook, and the flagged command execution is reached through the user-selected `cloop connect` workflow. Its detected downloads install coding agents from their official vendor sources, consistent with the package’s stated purpose.
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The package has no install, preinstall, or postinstall lifecycle hook; its scripts are build and development tasks.
- The executable dispatches `connect` to its connection workflow only when that command is selected.
- The connection workflow supports installing missing agent tools as part of setting up a Cloop machine.
- The Claude Code installer uses Anthropic’s official `claude.ai` installer endpoint.
Evidence against
- The connection workflow can run shell commands that download and execute the official Claude Code installer, and can install Codex during setup.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
5 flaggeddist/chunk-WC72IBTN.jsView file
16// src/commands/login.ts
L17: import { execFile } from "child_process";
L18:
High
Child Process
Package source references child process execution.
dist/chunk-WC72IBTN.jsView on unpkg · L16dist/chunk-G35UCZVP.jsView file
48windowsExecutable,
L49: windowsPowerShellEnv,
L50: writeBoxConfig
...
L107: import { basename, delimiter, dirname, join, posix, win32 } from "path";
L108: var CODEX_RELEASES = "https://github.[redacted]";
L109: var CODEX_TRIPLES = {
...
L127: var CODEX_SUMS = `${CODEX_RELEASES}/codex-package_SHA256SUMS`;
L128: function codexInstallCommand(home, platform = process.platform, arch = process.arch) {
L129: const asset = codexAsset(platform, arch);
...
L235: }
L236: function findLocalCodex(home, env = process.env) {
L237: findIn(codexBinDir(home), "codex", env);
High
Sandbox Evasion Gated Capability
Source gates dangerous network, credential, or execution behavior behind CI, host, platform, time, or geo fingerprint checks.
dist/chunk-G35UCZVP.jsView on unpkg · L48Findings
1 Critical3 High3 Medium5 Low
CriticalPrevious Version Dangerous Deltadist/chunk-LYHS7CSA.js
HighChild Processdist/chunk-WC72IBTN.js
HighSandbox Evasion Gated Capabilitydist/chunk-G35UCZVP.js
HighRuntime Package Installdist/chunk-G35UCZVP.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowSuspicious Lifecycle Evidencepackage.json
Affected versions and remediation
This report applies to @cloop/runner@0.3.37.
See all versions of @cloop/runner for other recorded verdicts.
- Review the evidence and your use of @cloop/runner@0.3.37 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.