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
19 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
The package has no install lifecycle hook, and the inspected flagged code supports documented CLI and MCP functions. Scanner hints were not substantiated as malicious behavior in source.
The inspected code reads agent and project data, writes a PID file using exclusive creation, and serves HTTP requests; no malicious persistence or unconsented agent control-surface mutation was established.
Decision evidence
public snapshot
Evidence for AI clean decision
package.json exposes CLI entrypoints and has no install lifecycle hook; its scripts begin with build tasks.
Agent registry scans project and user agent directories, while the skill registry reads command Markdown files.
The flagged helper files read Claude project data and project settings; these excerpts do not show an unconsented control-surface write.
The README diagram describes agents reading and writing the package database, not reviewer-directed instructions.
MCP server writes a PID file with exclusive creation and restrictive permissions; the UI server handles HTTP requests.
The generated agent router and design detector are described as package features; managed Chrome prefers an installed browser.
Evidence against
Scanner flags include persistence and agent-control-surface mutation, but inspected excerpts show PID bookkeeping and reads of agent settings or project data, not a backdoor or broad configuration write.
No inspected README match addresses or directs the security reviewer or scanner.
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.