Metadata-Version: 2.5
Name: skill-brain
Version: 2.1.0
Summary: Multi-agent engineering skill system for Claude Code, Codex CLI, Antigravity, Gemini CLI, and Cursor
License: MIT License
        
        Copyright (c) 2026 Skill Brain Contributors
        
        Permission is hereby granted, free of charge, to any person obtaining a copy
        of this software and associated documentation files (the "Software"), to deal
        in the Software without restriction, including without limitation the rights
        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
        copies of the Software, and to permit persons to whom the Software is
        furnished to do so, subject to the following conditions:
        
        The above copyright notice and this permission notice shall be included in all
        copies or substantial portions of the Software.
        
        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
        SOFTWARE.
License-File: LICENSE
Requires-Python: >=3.10
Provides-Extra: mcp
Description-Content-Type: text/markdown

# Skill Brain

> A pip-installable, multi-agent skill system for senior engineering discipline across Claude Code, Codex CLI, Antigravity, Gemini CLI, Cursor, and any AGENTS.md-aware AI coding agent.

[![Python Version](https://img.shields.io/badge/python-3.10%2B-blue.svg)](https://python.org)
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)

---

## What is Skill Brain?

Skill Brain transforms AI coding agents into disciplined engineering partners. Instead of producing generic AI templates, conversational filler, or unverified hallucinations, Skill Brain equips your agent with:

- **69 Precision Engineering Skills** covering UI design, backend architecture, TDD, debugging, typography, vector craftsmanship, and Git workflows.
- **Dedicated Pre-Implementation Planners** (`$plan-frontend`, `$plan-backend`, `$plan-debug`, `$plan-anti-slop`) that formulate blueprints before code is touched.
- **Universal Multi-Agent Compatibility**: Works natively with Claude Code (`.claude/skills/`), Codex CLI (`.agents/skills/`), Google Antigravity (`.agents/skills/`), Gemini CLI, and Cursor (`AGENTS.md` and MCP).
- **FastMCP Stdio Server**: Exposes skills, references, decision trees, and intelligent skill routing directly via the Model Context Protocol.
- **User-Controlled Virtual Memory**: Disk memory is strictly gated behind explicit `$memory` commands, keeping casual sessions clean.

---

## Quickstart

### 1. Installation

```bash
pip install skill-brain
```

With MCP server support:

```bash
pip install "skill-brain[mcp]"
```

### 2. Initialize in Any Project

Run inside any repository or workspace:

```bash
skill-brain init --agents all --mcp
```

This installs:
- `.agents/skills/<skill>/SKILL.md` (Open standard Agent Skills layout)
- `.agents/brain/` (decision matrices, config, and skill index)
- `.agents/memory/` (plans, phases, decisions, archive)
- Standard instruction files: `AGENTS.md`, `CLAUDE.md`, and `GEMINI.md`
- Project-level MCP configuration (`.mcp.json`, `.cursor/mcp.json`, `.gemini/settings.json`)

---

## Dollar Commands ($) & Triggers

Type any trigger in chat to activate specialized engineering suites:

| Command | Shorthand | Target Skill | Function |
|---|---|---|---|
| `$orchestrate` | `brain:` | `supreme-orchestrator` | Evaluates multi-domain intent, category routing, and confirmation gate |
| `$plan` | `plan:` | Domain Planners | Formulates task blueprint in chat with **zero disk writes** |
| `$plan-frontend` | `plan-ui:` | `frontend-planner` | Pre-build UI design & dual-viewport responsive blueprint |
| `$frontend` | `ui:`, `fe:` | `frontend-master` | Activates frontend suite with Taste Kill + Anti-SOP + Dynamic Sizing |
| `$plan-backend` | `plan-api:` | `backend-planner` | Domain schemas, responsive API contract, and TDD plan |
| `$backend` | `api:`, `be:` | `backend-master` | Karpathy simplicity, TDD loops, and responsive data contracts |
| `$plan-debug` | `plan-fix:` | `debug-planner` | Mandatory diagnostic symptom, hypothesis, and evidence plan |
| `$debug` | `fix:` | `debug-master` | Root-cause isolation, evidence collection before fixes |
| `$plan-anti-slop`| `audit-plan:` | `anti-slop-planner` | Scans for AI tells, produces 6-dimension remediation plan |
| `$anti-slop` | `gate:`, `audit:` | `anti-ai-sop-agent` | Standalone quality-gate scorecard (PASS / WARN / FAIL) |
| `$architect` | `arch:` | `architecture-master` | Worktree isolation, sub-agents, multi-session mapping |
| `$memory` | `memory:` | `memory-master` | Persistent disk-backed plans & execution phases with chat dashboard |
| `$taste` | `kill-slop:` | `design-taste-frontend`| 247-site study art direction: subject-derived palette & typography |
| `$typography` | `type:` | `typography-master` | Domain font archetypes (Swiss, Editorial, Mono, Humanist) |
| `$icon` | `svg:` | `svg-icon-craftsman` | Clean 24×24 grid-snapped vector geometry with 1.5px/2.0px stroke |
| `$verify` | `test:` | `test-verifier` | Demands raw CLI test/build output before claiming completion |
| `$git` | `git:` | `git-master` | Apex Git orchestrator: branch hygiene, atomic commits, phase sync |
| `$commit` | `commit:` | `git-atomic-committer` | Splits diffs into 2+ atomic commits across domains |
| `$branch` | `branch:` | `git-branch-manager` | Kebab branch taxonomy (`feat/`, `fix/`, `phase/`), rebase sync |
| `$phase-commit` | `phase-commit:`| `git-phase-committer` | Automatic memory phase commit hook with embedded verification proof |

---

## CLI Reference

- `skill-brain init [--agents all|claude,codex,antigravity,gemini,cursor] [--mcp] [--legacy-agent-dir]`
- `skill-brain update` — Refresh skills using hash manifest without overwriting user customizations
- `skill-brain doctor` — Comprehensive health check verifying frontmatter, sizes, and references
- `skill-brain list` — List all installed skills grouped by category
- `skill-brain show <name>` — Display full skill specification and references
- `skill-brain search <query>` — Keyword search across names, descriptions, and aliases
- `skill-brain add <path>` — Validate and install a custom skill package
- `skill-brain serve` — Run the FastMCP stdio server exposing skills to agent environments
- `skill-brain uninstall` — Safely remove Skill Brain components created by the manifest

---

## Configuration

`.agents/brain/config.json`:

```json
{
  "confirm": "always"
}
```

- `"always"` (default): Asks confirmation before reading or loading skills.
- `"multi-skill"`: Asks confirmation only when loading 2 or more skills simultaneously.
- `"never"`: Fully autonomous skill execution without prompt pauses.

---

## License

MIT © Skill Brain Contributors
