CLI engine for Monomind — an open-source MCP server for Claude Code, OpenCode, Antigravity, Kimi Code, and Codex, with a codebase knowledge graph (tree-sitter + SQLite), persistent memory, multi-agent task coordination, and session hooks. Apache 2.0 licen
Allowed — no malicious behavior detected
18 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
Source inspection found no npm install hook or concrete attack behavior. Scanner findings correspond to local daemon state, project helper behavior and user-invoked features, so the package is clean.
Use of the CLI or an associated feature, such as MCP daemon management, session hooks or browsing.
Impact
No confirmed attacker-controlled execution, credential disclosure or harmful control-surface mutation was established.
Mechanism
The inspected code creates local PID and helper state, scans agent definitions and sessions, and sends dashboard events to a local endpoint by default.
Decision evidence
public snapshot
Evidence for AI clean decision
package.json lists build, test and release scripts; no npm install lifecycle hook is present.
The agent-registry and skill-registry locations scan project or user definitions; the other required helper locations read session, project or settings data, or gate ADR generation on a setting.
The capture handler sends events to a dashboard URL defaulting to localhost and reads a local dashboard token; MCP PID state uses exclusive file creation.
The inspected CLI, design detector, local helper generators and browser fallback support invoked product features; README describes local state and identifies model-provider and feature-related outbound calls.
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.