Remote multi-agent workspace CLI with deep Claude integration, native terminal streaming, and self-hosting.
Allowed — no malicious behavior detected
23 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
The install hook extracts the package’s platform utilities and repairs node-pty permissions. Runtime credentials go to the configured Happy server for the CLI’s remote-workspace and push features, and the README clearly discloses that the server can access session contents and account secrets.
Postinstall unpacks the platform tools; runtime network activity occurs when the CLI or its daemon performs remote workspace and push operations.
Impact
No malicious impact identified. Users must trust their configured server with session contents and connected-account secrets, as the README states.
Mechanism
The API client supplies its configured server URL and credential to the push client. The README describes the server trust model and the data the server can access.
Evidence
tools/unpacked/*node-pty/prebuilds/*
Network endpoints2
veryhappy.devapi.cluster-fluster.com
Decision evidence
public snapshot
Evidence for AI clean decision
The postinstall hook only unpacks platform utility archives and fixes node-pty permissions; it does not run a downloader or unrelated command.
The API client passes its configured server URL to the push client, which sends the CLI credential to that same endpoint for push-token operations.
The README identifies the default server and explicitly warns that its operator can read session contents and account secrets.
The large CLI bundle’s decryption match is dotenv vault decryption; searches found no eval, Function construction, or URL-based payload execution in that bundle.
Evidence against
The CLI relays prompts, code, tool output, and account secrets to its server as part of its remote-workspace function; this is explicitly disclosed in the README.
A push-client constructor contains an api.cluster-fluster.com fallback, but the API client passes configuration.serverUrl when constructing it.
Behavioral surface
SourceChild process · Crypto · Dynamic require · Environment variables · Filesystem · Network · Shell · Web socket
Supply chainHigh entropy strings · Possible obfuscation · URL strings
ManifestNo manifest risk signals triggered.
Scanned 376 files · 7.38 MB of source
20 external domains referenced
127.0.0.1
accounts.google.com
api.cluster-fluster.com
api.openai.com
auth.openai.com
claude.ai
code.claude.com
console.anthropic.com
developers.google.com
docs.openclaw.ai
github.com
goo.gle
happy.mereith.com
host.example.ts.net
oauth2.googleapis.com
opencode.ai
registry.npmjs.org
veryhappy.dev
www.apple.com
www.googleapis.com
Findings
4 critical · 7 high · 7 medium · 5 low
Automated scan signals; the review checked them and cleared this version. Where they are, most severe file first. Code opens a file's flagged lines, with their evidence, below.
16import { c as claimSessionOrExit, b as AUTO_RESTARTED_ENV, R as RESTARTED_ON_REQUEST_ENV, a as takeAutoRestartedEvent, A as AUTO_RESTART_CAPABILITY } from'./sessionLock-ePizKIMs.m...
⋯
22 { B BaseReasoningProcessor, s settleCodexSession } ;