Ruflo - Enterprise AI agent orchestration for Claude Code. Deploy 60+ specialized agents in coordinated swarms with self-learning, fault-tolerant consensus, vector memory, and MCP integration
Allowed — no malicious behavior detected
20 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
The package has no npm install lifecycle hook; its only listed lifecycle hook is prepublishOnly. Inspected execution and configuration behavior is associated with explicit package features, and no concrete malicious behavior was found.
Static reason
One or more suspicious static signals were detected.
Trigger
A user runs project initialization, invokes a browser tool, or runs a benchmark.
Impact
No confirmed malicious effect or attack surface was established by these inspected behaviors.
Mechanism
The package creates or reads project-scoped helper and agent files, runs the Claude CLI for benchmark work, and invokes the browser automation CLI when requested.
Decision evidence
public snapshot
Evidence for AI clean decision
package.json defines a prepublishOnly hook, with no install lifecycle hook.
The hook handler resolves the package CLI from known package locations.
The intelligence helper searches project-local memory locations.
The statusline helper resolves the package CLI from known package locations.
The Codex initializer creates AGENTS.md during explicit project initialization and skips an existing file unless forced.
The helper generator reads project-local memory locations when generating helpers.
The guidance analyzer invokes the Claude CLI through execFile for a benchmark task.
The AgentBBS source describes MCP tools for its federation service.
Evidence against
Behavioral surface
SourceChild process · Crypto · Dynamic require · Environment variables · Filesystem · Network · Shell · Web socket
Supply chainHigh entropy strings · URL strings
ManifestNo manifest risk signals triggered.
Scanned 543 files · 8.13 MB of source
42 external domains referenced
100.104.125.72
127.0.0.1
agentbbs.local
aistudio.google.com
api.anthropic.com
api.cognitum.one
api.npmjs.org
api.openai.com
api.pinata.cloud
api.web3.storage
claude.com
cli.github.com
cloud.google.com
cloudflare-ipfs.com
cognitum.one
datasets-server.huggingface.co
dweb.link
funnel.ruv.io
gateway.pinata.cloud
generativelanguage.googleapis.com
git-scm.com
github.com
html.duckduckgo.com
hub.docker.com
huggingface.co
ipfs.io
json.schemastore.org
nodejs.org
ollama.com
openrouter.ai
pinata.cloud
qntm.org
raw.githubusercontent.com
registry.npmjs.org
relay.ruv.io
sql.js.org
storage.googleapis.com
us-central1-claude-flow.cloudfunctions.net
w3s.link
web3.storage
www.apple.com
x.ruv.io
Findings
7 high · 8 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.