Antigravity-First AI Tooling & Scaffolding Chassis

Modern Go Starter Engineered for AI Agents

Optimal developer defaults, pure-Go SQLite backlog DAG, zero-CGo static compilation, local LLM tiering ($0.00 first), and pluggable Model Context Protocol (MCP) server.

quick install
go install github.com/gosecio/gosec/cmd/gosec@latest
Get Started with gosec init Read Documentation

Core Pillars

Engineered for the Modern AI Agent Era

AI Architecture

Antigravity-First AI Tooling

Built ground-up for Google Antigravity (IDE, CLI agy, agent skills runtime). Includes progressive disclosure skills in .agents/skills and authoritative invariants.

Explore architecture →
Task Ledger

Pure-Go SQLite Backlog DAG

Embedded DAG task coordination in db/backlog/backlog.sqlite using modernc.org/sqlite. Zero CGo dependencies, strictly MCP-governed for AI coding agents.

Explore architecture →
FinOps

Local LLM Tiering ($0.00 First)

Seamlessly routes boilerplate generation and code analysis to local Ollama (qwen3-coder:30b / deepseek-r1:14b) with graceful fallback to Gemini 3.8 Flash.

Explore architecture →
Static Go

Zero-CGo Static Go

Compiled strictly with CGO_ENABLED=0 across Go 1.27.1. Hermetic, cross-platform static binaries with sub-millisecond execution.

Explore architecture →
Rule 14

Rule 14 Mandatory AI Coverage

Every capability is simultaneously delivered as an importable Go package, a CLI command (cmd/gosec), and a strongly typed stdio MCP tool.

Explore architecture →
Generator

Modular Scaffolding Engine

Interactive CLI wizard to scaffold Go archetypes (CLI, REST API, Stdio MCP Server, Library, Astro fullstack) with selectable features (Devcontainer, MkDocs, CI/CD).

Explore architecture →

Lifecycle Governance

The 6-Phase Clean-Room Engineering Cycle

Antigravity and AI agents follow strict invariants to ensure deterministic quality, zero regression, and reproducible static builds.

PHASE 01

Backlog

SQLite DAG stage transition & prerequisite checking strictly via stdio MCP tools.

PHASE 02

Research

Mandatory live web lookup before adopting external packages. Zero training-cutoff assumptions.

PHASE 03

TDD

Contract testing first. Race-condition-free Go concurrency verification (go test -race).

PHASE 04

Verification

Zero-CGo static binary compilation to bin/ and doctor system diagnostics.

PHASE 05

Documentation

Material for MkDocs, Mermaid diagrams, and LLM discoverability in docs/.

PHASE 06

Release

Multi-arch packaging, SHA256 checksum generation, and semantic version tagging.

Model Context Protocol

Multiplexed Stdio JSON-RPC 2.0 Server

cmd/mcp-server connects Antigravity and AI coding agents directly to your project's capabilities. Features are dynamically discovered and toggled in config.yaml.

  • Pure Go stdio JSON-RPC 2.0 multiplexer
  • Rule 14: Strong typing and input schema validation
  • Native Antigravity progressive disclosure integration
mcp_config.json → bin/mcp-server

backlog_get_ready_tasks

backlog

Calculates unblocked tasks whose DAG prerequisites are met in SQLite.

ollama_generate

ollama

Zero-cost local LLM inference via Ollama with automatic cloud fallback.

doctor_diagnose

doctor

Checks Go compiler version, Git status, and Ollama connectivity.

triage_log

triage

Analyzes CI failure logs and synthesizes root causes with Gemini API.

scaffold_package

scaffold

Scaffolds new Go package adhering to Rule 14 with Go API, MCP tool, & tests.

Getting Started

Zero-to-Production in Seconds

  1. Step 01

    Verify Go 1.27.1 compiler, Git VCS, and local Ollama inference server connectivity.

    $ gosec doctor
  2. Step 02

    Scaffold a new production-ready Go project with chosen components.

    $ gosec init myapp --archetype=cli
  3. Step 03

    Generate a new package adhering to Rule 14 with Go API, MCP tool, and unit tests.

    $ gosec new package audit --desc="Security audit trail"
  4. Step 04

    Compile zero-CGo static binaries and run race detector test suite.

    $ make build && make test

Connected Intelligence

The AI Engineering Ecosystem

Engineered for full-lifecycle AI software delivery: scaffolding, runtime security policy enforcement, and autonomous agent systems.

Scaffolding Chassis gosec.io

Go Starter & Task DAG

Zero-CGo static binaries, pure-Go SQLite backlog management, $0.00 local LLM tiering, and multiplexed stdio MCP tooling for Antigravity AI pair programming.

Read Docs → v0.0.2
Security Plane ctxsec.io

Context Security Plane

Declarative in-kernel hypervisor and hardware-enforced sandboxing for AI agents. Sub-5µs Google CEL security policies and context DLP protection.

Visit ctxsec.io → Hypervisor
Innovation Lab emaik.io

Autonomous AI Engineering

Digital innovation studio and systems lab researching cognitive AI agent workflows, edge developer infrastructure, and open-source ecosystems.

Discover emaik.io → Coded with 💖