@ggailabs/cli-context 0.5.6 → 1.0.1

package/README.md

@@ -1,195 +1,39 @@
-# @ggailabs/cli-context
+# 💎 @ggailabs/cli-context (Zenith v6.1)
 
-[![npm version](https://badge.fury.io/js/@ggailabs%2fcli-context.svg)](https://www.npmjs.com/package/@ggailabs/cli-context)
-[![CI](https://github.com/ggailabs/ggailabs-context/actions/workflows/ci.yml/badge.svg)](https://github.com/ggailabs/ggailabs-context/actions/workflows/ci.yml)
-[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+> **Genesis Grid AI Framework CLI** - The definitive universal orchestrator for AI-assisted development.
 
-A lightweight CLI that scaffolds living documentation and AI-agent playbooks for any repository—no LLMs or API keys required. The generated structure gives teams a consistent starting point for knowledge sharing while keeping everything under version control.
-
-## ⚙️ Requirements
-
-- Node.js 20+ (we currently test on 20, 22, 23, and 24)
-
-## ✨ What You Get
-
-- 📚 `docs/` folder with a documentation index plus ready-to-edit guides (overview, architecture, workflow, testing)
-- 🤖 `agents/` folder containing playbooks for common engineering agents and a handy index
-- 🔁 Repeatable scaffolding that you can re-run as the project evolves
-- 🧭 Repository-aware templates that highlight top-level directories for quick orientation
-- 🧠 AI-ready front matter and `agent-update` markers so assistants know exactly what to refresh
-
-## 📦 Installation
-
-Use `npx` to run the CLI without installing globally:
-
-```bash
-npx @ggailabs/cli-context
-```
-
-Or add it to your dev dependencies:
-
-```bash
-npm install --save-dev @ggailabs/cli-context
-```
+Orchestrate your codebase documentation, AI agent prompts, and development workflows under the rigor of the **Zenith Protocol v6.1**.
 
 ## 🚀 Quick Start
 
-```bash
-# Launch the interactive wizard
-npx @ggailabs/cli-context
-
-
-# Scaffold docs and agents into ./.context
-npx @ggailabs/cli-context init ./my-repo
-
-# Only generate docs
-npx @ggailabs/cli-context init ./my-repo docs
-
-# Only generate agent playbooks, with a custom output directory
-npx @ggailabs/cli-context init ./my-repo agents --output ./knowledge-base
-
-# Fill docs and agents with the repo context (preview the first 3 updates)
-npx @ggailabs/cli-context fill ./my-repo --output ./.context --limit 3
-
-# Draft a collaboration plan seeded with agent and doc touchpoints
-npx @ggailabs/cli-context plan release-readiness --output ./.context
-
-# Let the LLM enrich an existing plan with the latest context
-npx @ggailabs/cli-context plan release-readiness --output ./.context --fill --dry-run
-```
-
-> ℹ️ The CLI pings npm for fresh releases at startup. Set `AI_CONTEXT_DISABLE_UPDATE_CHECK=true` to skip the check.
-
-After running the command, inspect the generated structure:
-
-```
-.context/
-├── agents/
-│   ├── README.md
-│   ├── code-reviewer.md
-│   └── ...
-└── docs/
-    ├── README.md
-    ├── architecture.md
-    └── ...
-```
-
-Customize the Markdown files to reflect your project’s specifics and commit them alongside the code.
-
-## 🧠 Guided Updates for AI Assistants
-
-Need help filling in the scaffold? Use [`prompts/update_scaffold_prompt.md`](./prompts/update_scaffold_prompt.md) as the canonical instruction set for any LLM or CLI agent. It walks through:
-
-- Gathering repository context and locating `agent-update`/`agent-fill` markers.
-- Updating documentation sections while satisfying the YAML front matter criteria.
-- Aligning agent playbooks with the refreshed docs and recording evidence for maintainers.
-
-Share that prompt verbatim with your assistant to keep updates consistent across teams.
-
-### Available Doc Guides & Agent Types
-
-The scaffold includes the following guides and playbooks out of the box:
-
-- Docs: `project-overview`, `architecture`, `development-workflow`, `testing-strategy`, `glossary`, `data-flow`, `security`, `tooling`
-- Agents: `code-reviewer`, `bug-fixer`, `feature-developer`, `refactoring-specialist`, `test-writer`, `documentation-writer`, `performance-optimizer`, `security-auditor`, `backend-specialist`, `frontend-specialist`, `architect-specialist`
-
-### AI Marker Reference
-
-- `<!-- agent-update:start:section-id --> … <!-- agent-update:end -->` wrap the sections that AI assistants should rewrite with up-to-date project knowledge.
-- `<!-- agent-fill:slot-id --> … <!-- /agent-fill -->` highlight inline placeholders that must be replaced with concrete details before removing the wrapper.
-- `<!-- agent-readonly:context -->` flags guidance that should remain as-is; treat the adjacent content as instructions rather than editable prose.
-
-When contributing, focus edits inside `agent-update` regions or `agent-fill` placeholders and leave `agent-readonly` guidance untouched unless you have explicit maintainer approval.
-
-## 🛠 Commands
-
-### `init`
-Scaffold documentation and/or agent playbooks.
-
-```
-Usage: ggailabs-cli-context init <repo-path> [type]
-
-Arguments:
-  repo-path               Path to the repository you want to scan
-  type                    "docs", "agents", or "both" (default)
-
-Options:
-  -o, --output <dir>      Output directory (default: ./.context)
-  --exclude <patterns...> Glob patterns to skip during the scan
-  --include <patterns...> Glob patterns to explicitly include
-  -v, --verbose           Print detailed progress information
-  -h, --help              Display help for command
-```
-
-### `fill`
-Use an LLM to refresh scaffolded docs and agent playbooks automatically.
-
-```
-Usage: ggailabs-cli-context fill <repo-path>
-
-Options:
-  -o, --output <dir>      Scaffold directory containing docs/ and agents/ (default: ./.context)
-  -k, --api-key <key>     API key for the selected LLM provider
-  -m, --model <model>     LLM model to use (default: x-ai/grok-4-fast)
-  -p, --provider <name>   Provider (openrouter only)
-      --base-url <url>    Custom base URL for OpenRouter
-      --prompt <file>     Instruction prompt to follow (optional; uses bundled instructions when omitted)
-      --limit <number>    Maximum number of files to update in one run
-  -h, --help              Display help for command
-```
-
-Under the hood, the command loads the prompt above, iterates over every Markdown file in `.context/docs` and `.context/agents`, and asks the LLM to produce the fully updated content.
-
-### `plan`
-Create a collaboration plan that links documentation guides and agent playbooks, or fill an existing plan with LLM assistance.
-
-```
-Usage: ggailabs-cli-context plan <plan-name>
-
-Options:
-  -o, --output <dir>      Scaffold directory containing docs/ and agents/ (default: ./.context)
-      --title <title>     Custom title for the plan document
-      --summary <text>    Seed the plan with a short summary or goal statement
-  -f, --force             Overwrite the plan if it already exists (scaffold mode)
-      --fill              Use an LLM to fill or update the plan instead of scaffolding
-  -r, --repo <path>       Repository root to summarize for additional context (fill mode)
-  -k, --api-key <key>     API key for the selected LLM provider (fill mode)
-  -m, --model <model>     LLM model to use (default: x-ai/grok-4-fast)
-  -p, --provider <name>   Provider (openrouter only)
-      --base-url <url>    Custom base URL for OpenRouter
-      --prompt <file>     Instruction prompt to follow (optional; uses bundled instructions when omitted)
-      --dry-run           Preview changes without writing files
-      --include <patterns...>  Glob patterns to include during repository analysis
-      --exclude <patterns...>  Glob patterns to exclude from repository analysis
-  -h, --help              Display help for command
-```
-
-In scaffold mode the command creates `.context/plans/<plan-name>.md`, keeps a `plans/README.md` index, and reminds contributors to consult the agent handbook before delegating work to an AI assistant. In fill mode it will scaffold the plan automatically if it does not exist, then read the plan plus its referenced docs and agent playbooks, send that context to the LLM, and either preview or write the updated Markdown.
-
-💡 Tip: run `npx @ggailabs/cli-context` with no arguments to enter an interactive mode that guides you through scaffold and LLM-fill options.
-
-Prefer driving the update elsewhere? Just grab [`prompts/update_scaffold_prompt.md`](./prompts/update_scaffold_prompt.md) and run it in your favorite playground or agent host. When you’re ready to automate, drop your API key in `.env` (for example `OPENROUTER_API_KEY` and `OPENROUTER_MODEL`) and let `fill` handle the edits inline.
-
-## 🧰 Local Development
+Initialize the Zenith Protocol in any project:
 
 ```bash
-git clone https://github.com/ggailabs/ggailabs-context.git
-cd ggailabs-context
-npm install
-npm run build
-npm run test
+npx @ggailabs/cli-context init
 ```
 
-During development you can run the CLI directly against TypeScript sources:
+## 🧠 Key Features
 
-```bash
-npm run dev -- ./path/to/repo
-```
+- **Universal Agnostic Orchestration**: Automatically synchronizes context to **Gemini (Antigravity)**, **Windsurf (Cascade)**, **OpenAI Codex**, and **Cursor**.
+- **Deep Feed Discovery**: Interactive interrogation of project modules, objectives, and tech debt to eliminate AI hallucinations.
+- **Universal Skills**: Adheres to the `agentskills.io` standard, mapping project skills to `.agent/skills/`.
+- **Monitor Hub**: A real-time terminal dashboard to track context health, sync status, and FinOps metrics.
+- **GG-Methodology Integration**: Enforces the 7 Pillars of Excellence by Guilherme Giorgi (GG).
+
+## ⌨️ Commands
 
-## 🤝 Contributing
+- `init`: Initialize or upgrade a project to Zenith v6.1.
+- `monitor`: Open the Universal Context Hub & Health dashboard.
 
-See [`AGENTS.md`](./AGENTS.md) for contributor guidelines, coding standards, and release tips. Pull requests are welcome!
+## 🏛 Directory Structure
 
-## 📄 License
+The CLI manages the hidden `.context/` directory:
+- `docs/`: Architecture, standards, and mapping.
+- `agents/`: AI persona definitions.
+- `skills/`: Standardized executable workflows.
+- `plans/`: IA/Human collaboration design docs.
 
-MIT © Vinícius Lana
+---
+**Founder:** Guilherme Giorgi (GG)  
+**Org:** Genesis Grid AI Labs  
+*Engineering Excellence. Visual Perfection. Zero Defects.*

package/dist/.context/docs/GENESIS_SYSTEM_PROMPT.md

@@ -0,0 +1,62 @@
+# System Prompt: Genesis Grid Lead Architect
+
+You are the **Lead Frontend Architect** for **Genesis Grid AI Labs**. 
+Your mandate is to design and implement interfaces that strictly adhere to the **Genesis Grid Design System (v4.5)**.
+
+### 1. CORE IDENTITY & TONE
+- **Role:** You do not write generic UI code. You write Genesis-grade code: precise, consistent, scalable, accessible, and performance-oriented.
+- **Tone:** Confident, Technical, Precise, Minimal. No jokes or fluff.
+- **Branding:** 
+  - Official Name: "Genesis Grid AI Labs".
+  - Founder: Guilherme Giorgi.
+
+### 2. TECH STACK (STRICT)
+- **Framework:** React 19 + TypeScript.
+- **Bundler:** Vite.
+- **Styling:** Tailwind CSS (Class merging ONLY via `cn()`, variants via `cva()`).
+- **Icons:** **NEVER** import Lucide/FontAwesome directly. **ALWAYS** use:
+  `import { Icon } from "@/components/ui/icon";`
+
+### 3. TYPOGRAPHY CONTRACT
+Fonts are semantic, strictly defined by their purpose:
+- **Headings/UI Labels:** `font-sans` (Inter UI)
+- **Long Text/Descriptions:** `font-serif` (Source Serif 4)
+- **KPIs/Metrics/Financial:** `font-display` (Rajdhani) → *Always use `tracking-wide`*
+- **Tables/Numbers/Code:** `font-mono` (JetBrains Mono) → *Always use `tabular-nums`*
+
+### 4. DESIGN TOKENS & COLORS
+**ABSOLUTELY FORBIDDEN:** Hardcoded hex codes (`#000`, `#fff`), arbitrary spacing, or direct colors.
+**MANDATORY:** Use semantic tokens only.
+
+- **Surfaces:** `bg-background` (#0A0A0F), `bg-card` (#121217), `bg-popover`.
+- **Primary:** `text-primary` / `bg-primary` (Currently maps to **Gold** `#C8B298`).
+- **Brand Accents:**
+  - Indigo/Purple (`#5E5CE6`) and Green (`#30D158`) are reserved for theme accents.
+- **Primary Color Usage Rule:** Sparse usage (≤ 8% of UI). Use only for:
+  - Icons, Badges, Progress indicators, KPI numbers, Primary CTAs.
+  - **NEVER** for large backgrounds, full cards, or decorative borders.
+
+### 5. COMPONENT & LAYOUT ARCHITECTURE
+- **Primitives:** All utilized components (Button, Card, Badge) must exist in `@/components/ui/`. If missing, create one following existing patterns.
+- **Radius & Shape:**
+  - *Standard UI:* `rounded-xl` / `rounded-2xl`
+  - *Industrial/Data/Sales:* `rounded-sm` / `rounded-md`
+  - *Constraint:* Never mix radius styles in the same context.
+- **Navigation:** Never modify base Sidebar/Topbar. Register new entries via config.
+- **Responsive:** Mobile-first. Mandatory support for `md` and `lg` breakpoints.
+
+### 6. ENGINEERING DISCIPLINE & QUALITY
+- **Utilities:** Check existing utilities/components before writing new CSS.
+- **No Duplication:** No duplicated layouts or ad-hoc components outside the system.
+- **Accessibility:** Not optional.
+
+### 7. QUALITY ASSURANCE CHECKLIST
+Before outputting code, verify:
+1. [ ] Are imports using `@/` aliases?
+2. [ ] Are icons using the specific `Icon` wrapper component?
+3. [ ] Are all colors using Tailwind semantic tokens (e.g., `text-muted-foreground`)?
+4. [ ] Is the font usage correct (Rajdhani for numbers/KPIs, Inter for UI)?
+5. [ ] Are there zero hardcoded hex values?
+
+---
+**Deviation from these rules is considered a defect.**

package/dist/.context/eng/agents/architect.md

@@ -0,0 +1,15 @@
+# 🏛 Genesis Grid Architect (ENG)
+
+You are the **Lead Architect** of Genesis Grid AI Labs. Your mission is to ensure that the code follows the **7 Pillars of Excellence** and the **Modular Sovereignty** principle.
+
+## 🛡 Mandatory Protocols
+1.  **Modular Enforcer:** Never allow horizontal coupling between modules.
+2.  **Methodology Guardian:** Ensure every task passes through the **PREVICE** cycle.
+3.  **Strict Typing:** Reject any implementation that uses `any` or lacks proper TypeScript interfaces.
+
+## 🧠 Decision Logic
+- Before suggesting a move, check if the module definition is clear in `PROJECT_MAP.md`.
+- Always verify if the current context usage is below the 40% threshold.
+
+---
+*Engineering Excellence. Zero Defects.*