@melaya/runner@1.1.32
Run Melaya AI pipelines locally with your own LM Studio or Ollama models
AI Security Review
scanned 2mo ago · by lpm-firewall-aiNo install-time hook or covert import-time behavior was found. However, the explicit runner design gives its remote service broad code-execution and authenticated-account capabilities that remain unsafe to treat as clean.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The user runs the CLI with a runner token and the connected server sends a run or social-action request.
Impact
A compromised or abusive Melaya server could execute arbitrary Python on the runner host and act through the user's connected accounts.
Mechanism
Remote bundle and task delivery followed by local Python execution and credentialed web requests.
AI rationale
No install-time hook or covert import-time behavior was found. However, the explicit runner design gives its remote service broad code-execution and authenticated-account capabilities that remain unsafe to treat as clean.
Evidence
dist/sharedVendor.jsdist/connection.jsdist/linkedinProxy.jsdist/lumaLogin.js~/.melaya-runner/shared/~/.melaya-runner/agentscope/temporary melaya-run-<runId>/main.py~/.melaya-runner/linkedin/cookies.json
Network endpoints2
api.melaya.org/api/v1/runner/shared-bundlewww.linkedin.com/voyager/api
Decision evidence
public snapshotAI called this Suspicious at 93.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- The runner downloads a server-supplied Python bundle and writes its files into the user home cache.
- A server run request writes supplied Python into a temporary directory and executes it locally.
- The package sends local LinkedIn session cookies in requests whose path and method come from a remote request.
- The Luma login flow captures authenticated browser cookies and returns them through the cloud socket.
Evidence against
- package.json has no install, preinstall, or postinstall hook.
- These actions require an explicit CLI launch with a runner token and an active server connection.
- Browser authorization code denies cloud metadata and private-network destinations.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShellWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
8 flaggeddist/uvBootstrap.jsView file
20import { homedir, platform, arch } from "os";
L21: import { spawn, execSync } from "child_process";
L22: const CACHE_DIR = join(homedir(), ".melaya-runner");
High
Child Process
Package source references child process execution.
dist/uvBootstrap.jsView on unpkg · L20dist/codeWorker.jsView file
288// ulimit -t: hard CPU time in seconds.
L289: // We use /bin/sh -c "ulimit ...; exec sandbox-exec ..." so both limits
L290: // apply before the Node process starts. This is a soft-then-hard pair;
High
dist/browserProvisioner.jsView file
Findings
1 Critical6 High4 Medium6 Low
CriticalPrevious Version Dangerous Deltadist/browserProvisioner.js
HighChild Processdist/uvBootstrap.js
HighShelldist/codeWorker.js
HighSandbox Evasion Gated Capabilitydist/browserProvisioner.js
HighCloud Metadata Accessdist/browserAuthz.js
HighRuntime Package Installdist/lumaLogin.js
HighSemantic Analysis Limiteddist/linkedinLogin.js#virtual:normalized:round1
MediumNetwork
MediumEnvironment Vars
MediumShips Build Helperdist/localRagIngest.py
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
Affected versions and remediation
This report applies to @melaya/runner@1.1.32.
See all versions of @melaya/runner for other recorded verdicts.
- Review the evidence and your use of @melaya/runner@1.1.32 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.