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
20 low-signal patterns reviewed and cleared.
AI security review
scanned · by lpm-firewall-ai
The static hints point to agent integration and local process or filesystem capabilities, but inspected source places these within the package's CLI and session workflows. No lifecycle hook or concrete malicious behavior was established.
Static reason
One or more suspicious static signals were detected.
Trigger
Running the CLI's explicit initialization command or using its configured session helpers.
Impact
These capabilities affect local agent workflows, but inspected source does not establish credential theft, remote execution, or unconsented install-time mutation.
Mechanism
The inspected helpers read local agent directories, settings, and session project files to support registry, tracking, and coordination features.
Decision evidence
public snapshot
Evidence for AI clean decision
The manifest defines CLI entrypoints and has no preinstall, install, or postinstall hook.
The README describes Monomind as an agent coordination tool and documents its security layer; the reviewer-manipulation hint was not substantiated.
The registry and helper code targets project or user agent directories and Claude project session data.
The init command runs setup through an explicit `monomind init` action, and its post-init code discloses the package-aligned dependency install.
Evidence against
The inspected code supports project-local agent, session, and provider functionality; the observed paths do not establish credential exfiltration or an unconsented install-time attack.
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 926 files · 6.76 MB of source
37 external domains referenced
127.0.0.1
accounts.google.com
aider.chat
antigravity.google
api.anthropic.com
api.example.com
api.github.com
api.osv.dev
api.z.ai
cdnjs.cloudflare.com
cli.github.com
code.claude.com
code.visualstudio.com
cursor.com
docs.factory.ai
docs.github.com
docs.openclaw.ai
docs.x.ai
example.com
gemini.google.com
git-scm.com
github.com
hermes-agent.nousresearch.com
huggingface.co
kiro.dev
learn.chatgpt.com
monoes.me
nodejs.org
opencode.ai
raw.githubusercontent.com
registry.npmjs.org
services.nvd.nist.gov
www.apple.com
www.instagram.com
www.kimi.com
www.linkedin.com
x.com
Findings
8 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.