@agentskillkit/agent-skills 1.0.1

package/docs/PATTERNS_REFERENCE.md

@@ -0,0 +1,93 @@
+# Patterns Learned from everything-claude-code
+
+> Reference từ Anthropic hackathon winner repo
+
+---
+
+## 1. Package Manager Detection Pattern
+
+```javascript
+// Priority order for detection
+const DETECTION_PRIORITY = [
+    'CLAUDE_PACKAGE_MANAGER',           // 1. Env var
+    '.claude/package-manager.json',      // 2. Project config
+    'package.json.packageManager',       // 3. package.json field
+    'lock files',                        // 4. Lock file detection
+    '~/.claude/package-manager.json',    // 5. Global config
+    'first available'                    // 6. Fallback
+];
+```
+
+---
+
+## 2. Memory Persistence Hooks
+
+```json
+// Session lifecycle hooks
+{
+    "PreToolUse": "scripts/hooks/session-start.js",
+    "Stop": "scripts/hooks/session-end.js"
+}
+```
+
+**Pattern:** Save/load context across sessions automatically.
+
+---
+
+## 3. Continuous Learning
+
+```markdown
+/learn - Extract patterns mid-session into skills
+- Auto-extract reusable patterns
+- Store in knowledge base
+- Apply in future sessions
+```
+
+---
+
+## 4. Verification Loop
+
+```markdown
+/checkpoint - Save verification state
+/verify - Run verification loop
+
+1. Checkpoint vs continuous evals
+2. Grader types
+3. Pass@k metrics
+```
+
+---
+
+## 5. Strategic Compaction
+
+When context is full:
+1. Save current state
+2. Suggest manual compaction
+3. Preserve important context
+4. Resume seamlessly
+
+---
+
+## 6. Dynamic Contexts
+
+```
+contexts/
+├── dev.md      # Development mode
+├── review.md   # Code review mode
+├── research.md # Exploration mode
+```
+
+Inject different system prompts based on task type.
+
+---
+
+## Implementation Ideas for agent-skill-kit
+
+- [ ] Add package manager detection to CLI
+- [ ] Create memory persistence skill
+- [ ] Add /checkpoint command
+- [ ] Implement dynamic contexts
+
+---
+
+**Source:** https://github.com/affaan-m/everything-claude-code

package/docs/PLAN-consolidation-final.md

@@ -0,0 +1,14 @@
+# Final Consolidation Plan
+
+## 1. Goal
+Complete the integration by updating the skill registry and ensuring all new skills are discoverable.
+
+## 2. Steps
+1. Run `generate-registry.js` to index `git-conventions` and `code-reviewer`.
+2. Inspect `registry.json` to confirm presence.
+3. Clean up the `agent-skills` and `.agent` (root) folders if user agrees (will ask via notify_user).
+
+## 3. Deliverables
+- Fully populated `agentskillskit/.agent/skills`.
+- Updated Registry.
+- Zero legacy folders.

package/docs/PLAN-integration.md

@@ -0,0 +1,28 @@
+# Skill Integration Plan
+
+## 1. Goal
+Integrate remaining valuable skills from CoinPika legacy folders into the new `agentskillskit` architecture.
+
+## 2. Skills to Migrate
+
+### A. `git-conventions` (from `coinpika-commit-formatter`)
+- **Structure**:
+  - `SKILL.md` -> Standardize
+  - `scripts/commit_msg.py` -> Move to `scripts/` (python) or rewrite in Node (optional, but Python is fine if user has it).
+  - **Purpose**: Enforce Conventional Commits.
+
+### B. `code-reviewer` (from `coinpika-pr-reviewer`)
+- **Structure**:
+  - `SKILL.md` -> Standardize
+  - `scripts/review.js` -> Move to `scripts/`.
+  - **Purpose**: Automated PR quality checks.
+
+## 3. Execution Steps
+1. Create `.agent/skills/git-conventions` and `.agent/skills/code-reviewer`.
+2. Copy source files.
+3. Refactor `SKILL.md` to remove "CoinPika" branding and match new standard.
+4. Ensure scripts are executable via `npx my-agent-skills` or directly.
+
+## 4. Verification
+- Verify `git-conventions` by simulating a bad commit message.
+- Verify `code-reviewer` by running it on the current codebase.

package/docs/PLAN-rebrand-agent-skills-kit.md

@@ -0,0 +1,104 @@
+# Plan: Rebrand to Agent Skills Kit (Private License)
+
+> **Goal:** Rename entire project to "Agent Skills Kit", update domain to `agentskillskit.com`, and enforce a strict proprietary license.
+
+## 📋 Overview
+
+The project is undergoing a major rebranding and licensing shift. All public branding (Antigravity Kit, My Agent Skills) will be replaced with **Agent Skills Kit**. The open-source MIT license will be replaced with a **Proprietary Commercial License** to position the product as a premium "Big Tech" solution.
+
+## 🏗️ Project Type
+
+**TYPE: CLI / BACKEND Tooling**
+- Primary Agent: `backend-specialist` (handling CLI logic and file updates)
+- Secondary: `product-owner` (branding enforcement)
+
+## ✅ Success Criteria
+
+1.  **Project Name**: All references to "Antigravity Kit" / "My Agent Skills" replaced with "**Agent Skills Kit**".
+2.  **Website**: All URLs point to `agentskillskit.com`.
+3.  **License**: `LICENSE.md` contains strict proprietary terms (No redistribution, commercial use only).
+4.  **Verification**: CLI output displays new branding.
+
+## 🛠️ Tech Stack
+
+-   **Runtime**: Node.js (CLI), Python (Scripts)
+-   **Package Manager**: npm/pnpm
+-   **License Model**: Proprietary (Closed Source)
+
+## 📂 File Structure (Impacted Files)
+
+```plaintext
+agentskillskit/
+├── LICENSE              <-- [REPLACE] With Proprietary License
+├── package.json         <-- [UPDATE] Name, description, URL, license
+├── README.md            <-- [UPDATE] Branding, badges, links
+├── .agent/
+│   ├── ARCHITECTURE.md  <-- [UPDATE] Project name
+│   └── GEMINI.md        <-- [UPDATE] Project name
+└── packages/cli/        <-- [UPDATE] Internal strings
+```
+
+## 📝 Task Breakdown
+
+### Phase 1: Foundation (License & Meta)
+
+- [ ] **Create Proprietary License** `task_id: license_update`
+    -   **Agent**: `product-owner`
+    -   **Input**: "Big Tech" Private License text
+    -   **Output**: updated `LICENSE` file
+    -   **Verify**: Read file content matches proprietary terms.
+
+- [ ] **Update Package Metadata** `task_id: pkg_json_update`
+    -   **Agent**: `backend-specialist`
+    -   **File**: `package.json`, `packages/cli/package.json`
+    -   **Action**: Change name to `agent-skills-kit`, license to `UNLICENSED` (or private), homepage to `agentskillskit.com`.
+    -   **Verify**: `npm install` warns about private license (expected).
+
+### Phase 2: Branding Find & Replace
+
+- [ ] **Update Documentation Branding** `task_id: docs_rebrand`
+    -   **Agent**: `documentation-writer`
+    -   **Files**: `README.md`, `ARCHITECTURE.md`, `GEMINI.md`
+    -   **Action**: Replace "Antigravity Kit", "My Agent Skills" -> "Agent Skills Kit".
+    -   **Verify**: `grep` returns 0 matches for old names.
+
+- [ ] **Update Code Strings** `task_id: code_rebrand`
+    -   **Agent**: `backend-specialist`
+    -   **Files**: `packages/cli/bin/*.js`, `packages/cli/lib/*.js`
+    -   **Action**: Update CLI welcome messages and error/help text.
+    -   **Verify**: Run `ag-smart --help` shows "Agent Skills Kit".
+
+### Phase 3: Domain Update
+
+- [ ] **Update URLs** `task_id: url_update`
+    -   **Agent**: `backend-specialist`
+    -   **Action**: Replace `antigravity-kit.com` / `agentskillskit.com` (if any) -> `https://agentskillskit.com`.
+    -   **Verify**: Check links in `README.md` work (or point to new domain).
+
+## 🧪 Phase X: Verification
+
+### 1. Branding Check
+```bash
+# Must return NO results
+grep -r "Antigravity" .
+grep -r "My Agent Skills" .
+```
+
+### 2. License Check
+```bash
+# Must show Proprietary terms
+cat LICENSE
+# Must NOT show MIT
+```
+
+### 3. CLI Check
+```bash
+# Run the CLI
+node packages/cli/bin/ag-smart.js --version
+# Expect: "Agent Skills Kit v..."
+```
+
+## ✅ PHASE X COMPLETE
+- [ ] License is Private
+- [ ] Name is Agent Skills Kit
+- [ ] Domain is agentskillskit.com

package/docs/PLAN-refactor.md

@@ -0,0 +1,48 @@
+# Refactoring Plan: My Agent Skills (Standardization & Maintainability)
+
+## 1. Goal
+Biến `agentskillskit` từ một tập hợp "lai tạp" (Antigravity + Vercel + CoinPika) thành một hệ thống **Nhất quán (Consistent)**, **Module hóa (Modular)** và **Dễ mở rộng (Extensible)**.
+
+## 2. Current State (The Problem)
+- **Inconsistent Structure**:
+  - `governance`: Dùng `constitution/`, `doctrines/`.
+  - `frontend` (new): Dùng `rules/*.md` (Styles Vercel).
+  - `frontend-design` (old): Dùng single `SKILL.md`.
+  - `api-patterns`: Dùng python scripts.
+- **Scattered Logic**: Script nằm rải rác (`.agent/scripts`, skill scripts).
+- **Hard to Upgrade**: Nếu muốn thêm rule mới cho `frontend-design`, phải sửa file markdown lớn thay vì thêm 1 file nhỏ.
+
+## 3. The New Standard Structure (Proposal)
+Mỗi skill PHẢI tuân thủ cấu trúc chuẩn sau:
+
+```
+skill-name/
+├── SKILL.md            # Entry point (Router & Metadata)
+├── rules/              # [OPTIONAL] Atomic knowledge (Markdown)
+│   ├── rule-001.md
+│   └── rule-002.md
+├── scripts/            # [OPTIONAL] Executable logic (Node/Python)
+│   ├── audit.js
+│   └── fix.js
+└── tests/              # [OPTIONAL] Validation for the skill itself
+```
+
+## 4. Execution Phases
+
+### Phase 1: Core System Refactor (CLI & Shared)
+- **Action**: Move `ag-smart.js` and `.agent/scripts` into a proper `packages/cli` structure (giống Vercel repo gốc) hoặc folder `core/`.
+- **Benfit**: Tách biệt "Engine" (CLI) khỏi "Data" (Skills).
+
+### Phase 2: Unification (Skill Standardization)
+- **Action**: Convert các skill cũ (dạng Single File MD) sang dạng **Atomic Rules**.
+  - Ví dụ: Xé nhỏ `frontend-design/SKILL.md` thành `frontend-design/rules/colors.md`, `frontend-design/rules/typography.md`.
+- **Action**: Xóa bỏ các file rác hoặc legacy artifacts (ví dụ template code thừa).
+
+### Phase 3: Metadata & Indexing
+- **Action**: Tạo `registry.json` hoặc cơ chế auto-discovery để CLI biết có bao nhiêu skill và rule đang tồn tại mà không cần hardcode.
+
+## 5. Next Step
+Chúng ta sẽ bắt đầu với **Phase 1: Core System Refactor**.
+1. Tạo thư mục `core/cli`.
+2. Di chuyển `ag-smart.js` và `scripts/*.js` vào đó.
+3. Cập nhật `package.json` để trỏ đúng bin.

package/docs/PYTHON_SCRIPTS.md

@@ -0,0 +1,55 @@
+# Python Scripts Requirements
+
+This document lists the Python dependencies required for running the validation scripts in `.agent/skills/*/scripts/`.
+
+## Required Python Version
+
+- **Python 3.8+** recommended
+
+## Installation
+
+Most scripts use only Python standard library. However, some may require additional packages:
+
+```bash
+# Install common dependencies
+pip install pyyaml requests
+```
+
+## Available Scripts
+
+| Script | Location | Description |
+|--------|----------|-------------|
+| `security_scan.py` | vulnerability-scanner | OWASP security audit |
+| `lint_runner.py` | lint-and-validate | Run linting tools |
+| `schema_validator.py` | database-design | Validate DB schemas |
+| `ux_audit.py` | frontend-design | UX accessibility check |
+| `seo_checker.py` | seo-fundamentals | SEO validation |
+| `lighthouse_audit.py` | performance-profiling | Web performance audit |
+| `playwright_runner.py` | webapp-testing | E2E test runner |
+| `mobile_audit.py` | mobile-design | Mobile UI validation |
+| `i18n_checker.py` | i18n-localization | Translation check |
+| `geo_checker.py` | geo-fundamentals | GenAI optimization |
+| `api_validator.py` | api-patterns | API contract validation |
+| `test_runner.py` | testing-patterns | Test suite runner |
+| `type_coverage.py` | lint-and-validate | TypeScript coverage |
+| `accessibility_checker.py` | frontend-design | A11y audit |
+
+## Running Scripts
+
+```bash
+# Run from project root
+python .agent/skills/vulnerability-scanner/scripts/security_scan.py .
+
+# Or use npm scripts
+npm run scan:security
+npm run lint
+npm run checklist
+```
+
+## Troubleshooting
+
+If a script fails with import errors, ensure you have the required dependencies:
+
+```bash
+pip install -r requirements.txt  # If available in script directory
+```