antigravity-agents-core

mcp
Security Audit
Fail
Health Warn
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Fail
  • fs module — File system access in .agents/hooks/block-dangerous-git.cjs
  • execSync — Synchronous shell command execution in .agents/hooks/verify-on-stop.cjs
  • fs module — File system access in .agents/hooks/verify-on-stop.cjs
  • execSync — Synchronous shell command execution in .agents/hooks/verify-on-stop.js
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Universal agentic engineering layer for Google Antigravity. Features autonomous multi-agent orchestration, epistemic grounding, 9 strict verification gates, AST blast radius analysis, and target project adaptation.

README.md
AAC (Antigravity Agent Core) Banner

AAC (Antigravity Agent Core)

Minimalist autonomous engineering framework for Google Antigravity.
Engineered for Gemini 3.8 Flash (High) • Native Progressive Disclosure • Lifecycle Hooks • Multi-Agent Workspaces

Version 5.0.0 Platform Gemini 3.8 Flash 64 Skills License: MIT

Quick StartArchitectureRepository LayoutSkill CatalogMemory SystemLifecycle HooksMCP Setup


⚡ Overview

AAC (Antigravity Agent Core) is a minimalist autonomous engineering framework built for the Google Antigravity Customization Architecture. It provides focused pair programming, autonomous multi-agent task execution, and progressive context disclosure without token waste or hallucinated tooling.

🚀 Native Tool Calling

Seamless execution with Google Antigravity primitives: run_command, replace_file_content, view_file, and invoke_subagent.

🧠 Minimalist & Terse

Tailored for Gemini 3.8 Flash (High) via Caveman Protocol (fluff-free precision) and the 7-rung Ponytail Ladder (YAGNI to one-liners).

🔒 Zero Global Pollution

Strictly workspace-scoped within .agents/. Zero machine-global pollution (~/.gemini/config/) and zero package.json pollution.


⚡ Quick Start (Multi-Platform)

Install AAC into any existing project (Python, Go, Rust, Node, Java, PHP, C++) with a single command:

Option 1: Universal NPX (Any OS with Node.js)

# Via npm registry:
npx @rafaelghif/aac init

# Or directly from GitHub:
npx github:rafaelghif/antigravity-agents-core init

Option 2: Standalone Windows PowerShell (Zero Node.js Prerequisite)

irm https://raw.githubusercontent.com/rafaelghif/antigravity-agents-core/main/install.ps1 | iex

Option 3: Standalone Linux / macOS (Zero Node.js Prerequisite)

curl -fsSL https://raw.githubusercontent.com/rafaelghif/antigravity-agents-core/main/install.sh | bash

[!IMPORTANT]
Zero Package.json Pollution Guarantee: The installer will NEVER create, overwrite, or mutate package.json in your target repository. It cleanly scaffolds .agents/, AGENTS.md, and CONTEXT.md, and appends ignore rules to your .gitignore.


🏗️ Architecture Overview

The framework coordinates context precedence, lifecycle hooks, autonomous subagent graphs, and memory tiers:

flowchart TD
    User(["👤 User Request / Slash Command"]) --> Antigravity["Google Antigravity Engine (CLI / 2.0 / IDE)"]

    subgraph ContextHierarchy ["🧠 Context Precedence (AGENTS.md)"]
        AGENTS["1. AGENTS.md (Root Guidelines, ≤12k chars)"]
        Rules["2. .agents/rules/*.md (trigger: always_on)"]
        Hooks["3. .agents/hooks.json (PreToolUse, PostToolUse, Stop)"]
        Plugins["4. .agents/plugins/ (Packaged MCP & Sidecars)"]
        MCP["5. .agents/mcp_config.json (GitHub & Gitea Tools)"]
        Skills["6. .agents/skills/ (Progressive Disclosure)"]
    end

    Antigravity --> ContextHierarchy

    subgraph LifecycleEngine ["⚡ Lifecycle Event Hooks Engine"]
        PreTool["PreToolUse Hook (git-guardrails: blocks push / reset / clean)"]
        PostTool["PostToolUse Hook (automated linting & validation)"]
        StopHook["Stop Quality Gate (blocks exit on failing tests)"]
    end

    Antigravity --> LifecycleEngine

    subgraph SubagentsGraph ["🤖 Autonomous Subagents (invoke_subagent)"]
        Investigator["cavecrew-investigator (TypeName: research, Model: flash)"]
        Builder["cavecrew-builder (TypeName: self, Workspace: branch)"]
        Reviewer["cavecrew-reviewer / code-review (Two-Axis Reviewers)"]
    end

    Antigravity --> SubagentsGraph

    subgraph SidecarsEngine ["⚙️ Background Sidecars & Scheduled Tasks"]
        Sidecars["repo-health (Periodic git & branch hygiene monitor)"]
        CronSchedule["schedule tool (One-shot and recurring timers)"]
    end

    Antigravity --> SidecarsEngine

📂 Repository Layout

antigravity-agents/
├── assets/                            # Framework visual assets & documentation banners
│   └── banner.png                     # Vector branding banner
├── .agents/                           # Workspace configuration (Strictly isolated)
│   ├── hooks.json                     # Antigravity lifecycle hooks (PreToolUse, Stop)
│   ├── hooks/                         # Cross-platform hooks (block-dangerous-git.cjs, verify-on-stop.cjs)
│   ├── plugins.json                   # Explicit workspace plugin registration
│   ├── skills.json                    # Explicit workspace skills registration (64 skills)
│   ├── mcp_config.example.json        # Sanitized template for Gitea and GitHub MCP
│   ├── plugins/                       # Workspace plugins packaging tools & sidecars
│   │   └── workspace-integrations/    # Workspace integrations bundle
│   ├── rules/                         # Workspace-level rules with 'trigger: always_on'
│   │   ├── caveman.md                 # Ultra-compressed token communication protocol
│   │   ├── coding-standards.md        # SRP, fail-fast, and targeted replacement rules
│   │   ├── git-workflow.md            # Conventional Commits and atomic changes
│   │   ├── ponytail.md                # 7-rung minimalist code ladder (YAGNI to one-liners)
│   │   └── memory-management.md       # 5-tier memory hierarchy & cross-session protocol
│   └── skills/                        # 64 On-demand skills (Progressive disclosure)
├── bin/                               # Universal CLI executable
│   └── cli.mjs                        # Multi-platform installer (init, audit, doctor, list)
├── docs/                              # Framework documentation & decision records
│   ├── adr/                           # Architectural Decision Records (0001-memory, 0002-node-hooks)
│   ├── agents/                        # Skill specifications (issue-tracker, domain, triage-labels)
│   ├── templates/                     # Standardized session handoff template
│   └── audit-checklist-64-skills.md   # Persistent 8-dimension audit checklist for all 64 skills
├── tests/                             # Automated verification suites (node --test)
│   ├── memory-system.test.mjs         # Verified 64-skill criteria and memory architecture
│   └── cli.test.mjs                   # CLI tests (including zero-package.json pollution test)
├── install.ps1                        # Standalone Windows PowerShell 1-liner installer (Zero Node)
├── install.sh                         # Standalone Linux/macOS curl 1-liner installer (Zero Node)
├── package.json                       # Framework manifest for npm/npx distribution
├── .scratch/                          # Local session scratchpad & handoff staging (gitignored)
├── AGENTS.md                          # Root instructions unconditionally loaded per turn (<12k chars)
├── GEMINI.md                          # Pointer alias to AGENTS.md
└── CONTEXT.md                         # Living domain glossary & architectural boundaries

🎯 Autonomous Skills Suite

Skills are loaded into Antigravity via progressive disclosure: only names and descriptions are exposed in the initial context. When a user intent or slash command is matched, the agent reads the target SKILL.md directly via view_file.

1. Specification, Planning & Product (Click to expand)
Skill Description Primary File
ask-matt Meta-router recommending the best skill or workflow for any given task SKILL.md
grill-me Relentless interactive interview stress-testing plans, requirements, and edge cases SKILL.md
grill-with-docs Stress-tests a plan while simultaneously documenting ADRs and domain glossaries SKILL.md
to-spec Synthesizes conversations and codebase context into complete technical specifications SKILL.md
to-tickets Decomposes specs into vertical tracer-bullet tickets with dependency edges SKILL.md
wayfinder Maps large multi-session efforts into a map of frontier decisions under fog of war SKILL.md
to-questionnaire Turns decisions and missing requirements into structured questionnaires SKILL.md
prototype Builds throwaway prototypes to sanity check state models and user experience SKILL.md
2. Implementation, TDD & Verification (Click to expand)
Skill Description Primary File
tdd Test-driven development (Red → Green → Refactor) enforcing pre-agreed seams SKILL.md
implement Implements feature work from tickets or specifications using strict TDD SKILL.md
implement-spec Implements an entire spec using autonomous subagent task graphs SKILL.md
surgical-patch Surgical bug fixes at the narrowest responsible layer without scope creep SKILL.md
diagnosing-bugs Scientific hypothesis testing loop for intermittent bugs and regressions SKILL.md
code-review Two-axis parallel review checking Standards (Fowler smells) and Spec fidelity SKILL.md
safe-refactor Restructures code while preserving exact behavior with verification brackets SKILL.md
verify-and-stop Proves existing work meets acceptance conditions without expanding scope SKILL.md
3. Minimalist Code & Token Conservation (Click to expand)
Skill Description Primary File
ponytail Enforces the 7-rung minimalist code ladder (YAGNI to one-liners) SKILL.md
ponytail-audit Whole-repo audit scanning for dead code, unneeded dependencies, and bloat SKILL.md
ponytail-debt Harvests ponytail: comments into a prioritized technical debt ledger SKILL.md
ponytail-review Code review focused exclusively on eliminating over-engineering SKILL.md
caveman Ultra-compressed token communication protocol preserving 100% technical accuracy SKILL.md
cavecrew Subagent delegation protocol with compressed output contracts to save context SKILL.md
caveman-stats Calculates real session token usage, turn count, and token savings SKILL.md
caveman-explore Read-only repository explorer returning compact path:line citations SKILL.md
4. Architecture, Governance & Safety (Click to expand)
Skill Description Primary File
git-guardrails Sets up PreToolUse hooks to intercept and block destructive git commands SKILL.md
handoff Checkpoints current conversation into .scratch/handoff.md for seamless resumption SKILL.md
subagent-handoff Hands off conversation state to an autonomous background worker SKILL.md
domain-modeling Builds and sharpens domain boundaries, CONTEXT.md, and ADRs SKILL.md
triage Triages incoming issues and PRs through canonical 5-state lifecycle SKILL.md
setup-ts-deep-modules Configures dependency-cruiser to enforce deep module boundaries SKILL.md
retro Post-session retrospective identifying environment and prompt improvements SKILL.md

🧠 5-Tier Memory Management

To maintain crisp context without attention degradation or token bloat, AAC partitions memory across five distinct tiers:

┌─────────────────────────────────────────────────────────────┐
│ Tier 1: Ephemeral Working Context (Intra-Session)           │ -> Context window & transcript.jsonl
├─────────────────────────────────────────────────────────────┤
│ Tier 2: Workspace Directives (Cross-Session Deterministic)  │ -> AGENTS.md (<12k chars), .agents/rules/*.md
├─────────────────────────────────────────────────────────────┤
│ Tier 3: Domain & Architectural Knowledge (Living Docs)      │ -> CONTEXT.md, docs/adr/*.md
├─────────────────────────────────────────────────────────────┤
│ Tier 4: Session Bridge Handoff (Inter-Session State)        │ -> .scratch/handoff.md via handoff skill
├─────────────────────────────────────────────────────────────┤
│ Tier 5: External Task Graph (Durable Frontier)              │ -> GitHub / Gitea Issues (to-tickets)
└─────────────────────────────────────────────────────────────┘
  1. Intra-Session (Tier 1): Ephemeral working context managed via progressive disclosure.
  2. Workspace Directives (Tier 2): Core guidelines in AGENTS.md and rules in .agents/rules/ (trigger: always_on).
  3. Domain Knowledge (Tier 3): Living domain glossary in CONTEXT.md and immutable decisions in docs/adr/.
  4. Session Bridge (Tier 4): Cold-start checkpoint saved to .scratch/handoff.md before exit. Rehydrated in new sessions via @handoff.md.
  5. Durable Task Graph (Tier 5): External source of truth for work items managed via GitHub or Gitea issues.

🛡️ Lifecycle Hooks & Security

AAC integrates with the native Antigravity lifecycle hook engine configured in .agents/hooks.json:

1. Git Guardrails (PreToolUse)

Blocks destructive git commands before run_command executes:

# Cross-platform Node.js hook (Windows, macOS, Linux):
echo '{"toolCall":{"name":"run_command","args":{"CommandLine":"git push origin main"}}}' | node .agents/hooks/block-dangerous-git.cjs
# Output: {"decision":"deny","reason":"BLOCKED: 'git push origin main' matches dangerous git pattern '\\bgit\\s+push\\b'..."}

2. Quality Gate (Stop)

Prevents terminating an agent session if automated tests fail:

{
  "quality-gate": {
    "enabled": true,
    "Stop": [
      {
        "type": "command",
        "command": "node hooks/verify-on-stop.cjs",
        "timeout": 15
      }
    ]
  }
}

🔌 Model Context Protocol (MCP)

AAC natively supports workspace-scoped MCP servers with credential sandboxing.

Copy .agents/mcp_config.example.json to .agents/mcp_config.json (gitignored):

cp .agents/mcp_config.example.json .agents/mcp_config.json

Configure local Gitea (stdio) and remote GitHub (SSE) connections:

{
  "mcpServers": {
    "github": {
      "serverUrl": "https://api.githubcopilot.com/mcp/",
      "headers": {
        "Authorization": "Bearer YOUR_GITHUB_PAT"
      }
    },
    "gitea": {
      "command": "gitea-mcp",
      "args": ["-t", "stdio"],
      "env": {
        "GITEA_HOST": "https://gitea.com",
        "GITEA_ACCESS_TOKEN": "YOUR_GITEA_PAT"
      }
    }
  }
}

🧪 Verification & Testing

Every commit and installer is validated against a comprehensive automated test suite:

npm test
✔ frontmatter has valid Antigravity name and description
✔ description specifies what it does and when to invoke
✔ body specifies Antigravity research subagent and native tools
✔ artifact carries no placeholder markers
✔ CLI --version prints v5.0.0
✔ CLI --help prints usage banner
✔ CLI list displays skills count
✔ CLI doctor performs environment health checks
✔ CLI init never creates or overwrites package.json in target directory
✔ AGENTS.md remains strictly below 12000 characters limit
✔ memory-management rule exists with trigger: always_on
✔ CONTEXT.md living domain document exists at root
✔ ADR 0001 records 5-tier memory decision
✔ docs/agents configuration files exist and are populated
✔ gitignore correctly ignores .scratch contents and preserves .gitkeep
✔ session handoff template exists
✔ all 64 skills comply with Antigravity operational criteria
ℹ pass 17, fail 0, duration_ms 443

📜 License

Distributed under the MIT License. See LICENSE for more information.

Developed by Muhammad Rafael Ghifari (@rafaelghif) for the Google Antigravity Ecosystem.

Reviews (0)

No results found