antigravity-ide 4.2.0 → 4.2.1

package/.agent/workflows/orchestrate.md

@@ -1,50 +1,237 @@
 ---
-description: Task quá chua? Gọi cả hội đồng chuyên gia vào làm theo quy chuẩn Senior.
+description: Task quá chua? Gọi cả hội đồng chuyên gia vào làm.
 ---
 
-# /orchestrate - Multi-Agent Command System
+# Multi-Agent Orchestration
 
+You are now in **ORCHESTRATION MODE**. Your task: coordinate specialized agents to solve this complex problem.
+
+## Task to Orchestrate
 $ARGUMENTS
 
 ---
 
-## 🔴 THE CONSTITUTION OF ORCHESTRATION
-1. **Min 3 Agents**: Single-agent delegation is NOT orchestration.
-2. **Context Passing**: Every subagent MUST receive the original intent + previous agent's results.
-3. **Phase Separation**: Planning MUST happen before Parallel Execution.
+## 🔴 CRITICAL: Minimum Agent Requirement
+
+> ⚠️ **ORCHESTRATION = MINIMUM 3 DIFFERENT AGENTS**
+> 
+> If you use fewer than 3 agents, you are NOT orchestrating - you're just delegating.
+> 
+> **Validation before completion:**
+> - Count invoked agents
+> - If `agent_count < 3` → STOP and invoke more agents
+> - Single agent = FAILURE of orchestration
+
+### Agent Selection Matrix
+
+| Task Type | REQUIRED Agents (minimum) |
+|-----------|---------------------------|
+| **Web App** | frontend-specialist, backend-specialist, test-engineer |
+| **API** | backend-specialist, security-auditor, test-engineer |
+| **UI/Design** | frontend-specialist, seo-specialist, performance-optimizer |
+| **Database** | database-architect, backend-specialist, security-auditor |
+| **Full Stack** | project-planner, frontend-specialist, backend-specialist, devops-engineer |
+| **Debug** | debugger, explorer-agent, test-engineer |
+| **Security** | security-auditor, penetration-tester, devops-engineer |
+
+---
+
+## Pre-Flight: Mode Check
+
+| Current Mode | Task Type | Action |
+|--------------|-----------|--------|
+| **plan** | Any | ✅ Proceed with planning-first approach |
+| **edit** | Simple execution | ✅ Proceed directly |
+| **edit** | Complex/multi-file | ⚠️ Ask: "This task requires planning. Switch to plan mode?" |
+| **ask** | Any | ⚠️ Ask: "Ready to orchestrate. Switch to edit or plan mode?" |
+
+---
+
+## 🔴 STRICT 2-PHASE ORCHESTRATION
+
+### PHASE 1: PLANNING (Sequential - NO parallel agents)
+
+| Step | Agent | Action |
+|------|-------|--------|
+| 1 | `project-planner` | Create docs/PLAN.md |
+| 2 | (optional) `explorer-agent` | Codebase discovery if needed |
+
+> 🔴 **NO OTHER AGENTS during planning!** Only project-planner and explorer-agent.
+
+### ⏸️ CHECKPOINT: User Approval
+
+```
+After PLAN.md is complete, ASK:
+
+"✅ Plan oluşturuldu: docs/PLAN.md
+
+Onaylıyor musunuz? (Y/N)
+- Y: Implementation başlatılır
+- N: Planı düzeltirim"
+```
+
+> 🔴 **DO NOT proceed to Phase 2 without explicit user approval!**
+
+### PHASE 2: IMPLEMENTATION (Parallel agents after approval)
+
+| Parallel Group | Agents |
+|----------------|--------|
+| Foundation | `database-architect`, `security-auditor` |
+| Core | `backend-specialist`, `frontend-specialist` |
+| Polish | `test-engineer`, `devops-engineer` |
+
+> ✅ After user approval, invoke multiple agents in PARALLEL.
+
+## Available Agents (17 total)
+
+| Agent | Domain | Use When |
+|-------|--------|----------|
+| `project-planner` | Planning | Task breakdown, PLAN.md |
+| `explorer-agent` | Discovery | Codebase mapping |
+| `frontend-specialist` | UI/UX | React, Vue, CSS, HTML |
+| `backend-specialist` | Server | API, Node.js, Python |
+| `database-architect` | Data | SQL, NoSQL, Schema |
+| `security-auditor` | Security | Vulnerabilities, Auth |
+| `penetration-tester` | Security | Active testing |
+| `test-engineer` | Testing | Unit, E2E, Coverage |
+| `devops-engineer` | Ops | CI/CD, Docker, Deploy |
+| `mobile-developer` | Mobile | React Native, Flutter |
+| `performance-optimizer` | Speed | Lighthouse, Profiling |
+| `seo-specialist` | SEO | Meta, Schema, Rankings |
+| `documentation-writer` | Docs | README, API docs |
+| `debugger` | Debug | Error analysis |
+| `game-developer` | Games | Unity, Godot |
+| `orchestrator` | Meta | Coordination |
+
+---
+
+## Orchestration Protocol
+
+### Step 1: Analyze Task Domains
+Identify ALL domains this task touches:
+```
+□ Security     → security-auditor, penetration-tester
+□ Backend/API  → backend-specialist
+□ Frontend/UI  → frontend-specialist
+□ Database     → database-architect
+□ Testing      → test-engineer
+□ DevOps       → devops-engineer
+□ Mobile       → mobile-developer
+□ Performance  → performance-optimizer
+□ SEO          → seo-specialist
+□ Planning     → project-planner
+```
+
+### Step 2: Phase Detection
+
+| If Plan Exists | Action |
+|----------------|--------|
+| NO `docs/PLAN.md` | → Go to PHASE 1 (planning only) |
+| YES `docs/PLAN.md` + user approved | → Go to PHASE 2 (implementation) |
+
+### Step 3: Execute Based on Phase
+
+**PHASE 1 (Planning):**
+```
+Use the project-planner agent to create PLAN.md
+→ STOP after plan is created
+→ ASK user for approval
+```
+
+**PHASE 2 (Implementation - after approval):**
+```
+Invoke agents in PARALLEL:
+Use the frontend-specialist agent to [task]
+Use the backend-specialist agent to [task]
+Use the test-engineer agent to [task]
+```
+
+**🔴 CRITICAL: Context Passing (MANDATORY)**
+
+When invoking ANY subagent, you MUST include:
+
+1. **Original User Request:** Full text of what user asked
+2. **Decisions Made:** All user answers to Socratic questions
+3. **Previous Agent Work:** Summary of what previous agents did
+4. **Current Plan State:** If plan files exist in workspace, include them
+
+**Example with FULL context:**
+```
+Use the project-planner agent to create PLAN.md:
+
+**CONTEXT:**
+- User Request: "Öğrenciler için sosyal platform, mock data ile"
+- Decisions: Tech=Vue 3, Layout=Grid Widget, Auth=Mock, Design=Genç Dinamik
+- Previous Work: Orchestrator asked 6 questions, user chose all options
+- Current Plan: playful-roaming-dream.md exists in workspace with initial structure
+
+**TASK:** Create detailed PLAN.md based on ABOVE decisions. Do NOT infer from folder name.
+```
+
+> ⚠️ **VIOLATION:** Invoking subagent without full context = subagent will make wrong assumptions!
+
+
+### Step 4: Verification (MANDATORY)
+The LAST agent must run appropriate verification scripts:
+```bash
+python .agent/skills/vulnerability-scanner/scripts/security_scan.py .
+python .agent/skills/lint-and-validate/scripts/lint_runner.py .
+```
+
+### Step 5: Synthesize Results
+Combine all agent outputs into unified report.
 
 ---
 
-## 🟢 PHASE 1: Strategic Planning (Sequential)
-**Agent**: `project-planner`
-- **Task**: Create `PLAN.md`.
-- **Gate**: Stop and ask for User Approval. **DO NOT PROCEED** without a "YES".
+## Output Format
+
+```markdown
+## 🎼 Orchestration Report
+
+### Task
+[Original task summary]
+
+### Mode
+[Current Antigravity Agent mode: plan/edit/ask]
 
-## 🔵 PHASE 2: Parallel Execution
-**Agent**: `orchestrator`
-- **Execution Matrix**:
-  - **Group A (Foundation)**: `database-architect`, `security-auditor`.
-  - **Group B (Core)**: `backend-specialist`, `frontend-specialist`.
-  - **Group C (Optimization)**: `performance-optimizer`, `seo-specialist`.
+### Agents Invoked (MINIMUM 3)
+| # | Agent | Focus Area | Status |
+|---|-------|------------|--------|
+| 1 | project-planner | Task breakdown | ✅ |
+| 2 | frontend-specialist | UI implementation | ✅ |
+| 3 | test-engineer | Verification scripts | ✅ |
 
-## 🔴 PHASE 3: Systemic Review
-**Agent**: `quality-inspector`
-- **Task**: Verify the convergence of all agents.
-- **Automation**: Run `security_scan.sh` and `lint_check.sh`.
-- **Handoff**: Create a unified `walkthrough.md`.
+### Verification Scripts Executed
+- [x] security_scan.py → Pass/Fail
+- [x] lint_runner.py → Pass/Fail
+
+### Key Findings
+1. **[Agent 1]**: Finding
+2. **[Agent 2]**: Finding
+3. **[Agent 3]**: Finding
+
+### Deliverables
+- [ ] PLAN.md created
+- [ ] Code implemented
+- [ ] Tests passing
+- [ ] Scripts verified
+
+### Summary
+[One paragraph synthesis of all agent work]
+```
 
 ---
 
-## Reporting Format:
-- **Orchestration Report**:
-  - Summary of Task.
-  - Agents invoked + specific contributions.
-  - Verification results.
-  - Deliverable links.
+## 🔴 EXIT GATE
+
+Before completing orchestration, verify:
+
+1. ✅ **Agent Count:** `invoked_agents >= 3`
+2. ✅ **Scripts Executed:** At least `security_scan.py` ran
+3. ✅ **Report Generated:** Orchestration Report with all agents listed
+
+> **If any check fails → DO NOT mark orchestration complete. Invoke more agents or run scripts.**
 
 ---
 
-## Critical Failure Modes (REJECT IF):
-- `agent_count < 3`.
-- No `walkthrough.md` created at the end.
-- Agents working on the same file without a split-state strategy.
+**Begin orchestration now. Select 3+ agents, execute sequentially, run verification scripts, synthesize results.**

package/.agent/workflows/performance-max.md

@@ -0,0 +1,9 @@
+---
+description: Optimize application performance to reach perfect Core Web Vitals.
+---
+
+1. Run Lighthouse/PageSpeed tests via browser tool.
+2. Use `performance-optimizer` to identify bottlenecks.
+3. Use `frontend-specialist` to optimize assets and code splitting.
+4. Use `backend-specialist` to optimize database queries and caching.
+5. Use `growth-engineer` to verify the impact on SEO metrics.

package/.agent/workflows/plan.md

@@ -1,58 +1,89 @@
 ---
-description: Chưa biết bắt đầu từ đâu? Lập kế hoạch theo chuẩn Senior Personnel.
+description: Chưa biết bắt đầu từ đâu? Lập kế hoạch trước.
 ---
 
-# /plan - Strategic Planning System
+# /plan - Project Planning Mode
 
 $ARGUMENTS
 
 ---
 
-## 🟢 PHASE 1: Discovery & Terrain Analysis (Gatekeeper)
-**Agent**: `explorer-agent`
-**Mission**: Understand the current state before proposing changes.
-- **Action**: Perform a recursive scan of the workspace.
-- **Action**: Identify all relevant DNA (`GEMINI.md`) and Rules (`rules/`).
-- **Critical Gate**: If the request is ambiguous, trigger the **Socratic Gate** and ask 3-5 clarifying questions.
+## 🔴 CRITICAL RULES
 
-## 🟡 PHASE 2: Strategic Implementation Plan
-**Agent**: `project-planner`
-**Mission**: Create the blueprint for success.
-- **Output**: `PLAN-{task-slug}.md` in the project root or relevant docs folder.
-- **Requirement**: Use GitHub-style alerts (IMPORTANT/WARNING) for risks.
-- **Protocol**: 
-  1. Define Clear Goals.
-  2. Map Dependency Chains.
-  3. Establish Phase-by-Phase Breakdown.
-  4. Create a specific Verification Plan (Automated + Manual).
+1. **NO CODE WRITING** - This command creates plan file only
+2. **Use project-planner agent** - NOT Antigravity Agent's native Plan mode
+3. **Socratic Gate** - Ask clarifying questions before planning
+4. **Dynamic Naming** - Plan file named based on task
 
-## 🔵 PHASE 3: Surgical Task Distribution
-**Agent**: `orchestrator`
-**Mission**: Map the plan to specialized specialists.
-- **Action**: Update `task.md` with unique IDs for every step.
-- **Action**: Assign the "Heavy Lifters" (Backend, Frontend, etc.).
+---
+
+## Task
+
+Use the `project-planner` agent with this context:
+
+```
+CONTEXT:
+- User Request: $ARGUMENTS
+- Mode: PLANNING ONLY (no code)
+- Output: docs/PLAN-{task-slug}.md (dynamic naming)
+
+NAMING RULES:
+1. Extract 2-3 key words from request
+2. Lowercase, hyphen-separated
+3. Max 30 characters
+4. Example: "e-commerce cart" → PLAN-ecommerce-cart.md
+
+RULES:
+1. Follow project-planner.md Phase -1 (Context Check)
+2. Follow project-planner.md Phase 0 (Socratic Gate)
+3. Create PLAN-{slug}.md with task breakdown
+4. DO NOT write any code files
+5. REPORT the exact file name created
+```
+
+---
 
-## 🔴 PHASE 4: Plan Validation & Sign-off
-**Agent**: `quality-inspector`
-**Mission**: Ensure the plan is "Operational-Ready."
-- **Verification**: Check if the plan matches `DNA_REF` compliance.
-- **Reporting**: Report the exact file path of the created Plan to the User.
+## Expected Output
+
+| Deliverable | Location |
+|-------------|----------|
+| Project Plan | `docs/PLAN-{task-slug}.md` |
+| Task Breakdown | Inside plan file |
+| Agent Assignments | Inside plan file |
+| Verification Checklist | Phase X in plan file |
 
 ---
 
-## Output Format
-```markdown
-[OK] Plan Created: {path_to_plan.md}
+## After Planning
+
+Tell user:
+```
+[OK] Plan created: docs/PLAN-{slug}.md
 
-### Next Steps:
-1. Review the Plan.
-2. Manually adjust if needed.
-3. Run `/create` or `/orchestrate` to begin the surgical execution.
+Next steps:
+- Review the plan
+- Run `/create` to start implementation
+- Or modify plan manually
 ```
 
 ---
 
-## Key Principles:
-- **No Code**: This workflow is strictly for strategy.
-- **Naming Protocol**: `PLAN-kebab-case-slug.md`.
-- **User-Centric**: Respect the user's OS and workspace constraints.
+## Naming Examples
+
+| Request | Plan File |
+|---------|-----------|
+| `/plan e-commerce site with cart` | `docs/PLAN-ecommerce-cart.md` |
+| `/plan mobile app for fitness` | `docs/PLAN-fitness-app.md` |
+| `/plan add dark mode feature` | `docs/PLAN-dark-mode.md` |
+| `/plan fix authentication bug` | `docs/PLAN-auth-fix.md` |
+| `/plan SaaS dashboard` | `docs/PLAN-saas-dashboard.md` |
+
+---
+
+## Usage
+
+```
+/plan e-commerce site with cart
+/plan mobile app for fitness tracking
+/plan SaaS dashboard with analytics
+```

package/.agent/workflows/security.md

@@ -1,47 +1,26 @@
 ---
-description: Sợ bị hack? Quét lỗ hổng và bảo mật ngay theo chuẩn Security Senior.
+description: Sợ bị hack? Quét lỗ hổng và bảo mật ngay.
 ---
 
-# /security - Professional Armor & Security Audit
+# /security - Security Hardening
 
 $ARGUMENTS
 
 ---
 
-## 🟢 PHASE 1: Attack Surface Mapping
-**Agent**: `penetration-tester` & `explorer-agent`
-**Mission**: Find where the system is "thin."
-- **Action**: Identify all public endpoints, input fields, and storage locations.
-- **DNA Link**: Check `rules/malware-protection.md` for external link risks.
+## Task
+Focus specifically on securing the application.
 
-## 🟡 PHASE 2: Vulnerability Research & Tooling
-**Agent**: `security-auditor`
-**Mission**: Run the deep scans.
-- **Action**: Run SAST/DAST tools (e.g., `npm audit`, `snyk`, `owasp-zap`).
-- **Research**: Check for common library vulnerabilities (CVEs).
-
-## 🔵 PHASE 3: Surgical Hardening
-**Agent**: `backend-specialist` & `devops-engineer`
-**Mission**: Patch the leaks.
-- **Action**: Implement Rate Limiting, Input Sanitization, and CSP headers.
-- **Protocol**: Apply "Least Privilege" to all IAM and system roles.
-
-## 🔴 PHASE 4: Verification & Compliance Report
-**Agent**: `quality-inspector`
-**Mission**: Confirm the "Fortress" is secure.
-- **Verification**: Re-run the exploit script to ensure it's blocked.
-- **Artifact**: Create a "Security Risk Table" in the `walkthrough.md`.
-
----
-
-## Security Mandates:
-- **No Hardcoding**: Reject any plan that hardcodes a secret.
-- **Sanitize Everything**: All user input is untrusted by default.
-- **Zero Trust**: Authentication must be verified at every layer.
+### Steps:
+1.  **Dependency Analysis**: Scan `package.json` for weak deps.
+2.  **Secret Detection**: Scan code for hardcoded API keys.
+3.  **Headers Check**: Verify Helmet/CORS configuration.
+4.  **Auth Review**: Check JWT/Session implementation logic.
 
 ---
 
-## Examples:
-- `/security audit all endpoints`
-- `/security scan for data leaks in logs`
-- `/security harden auth flow`
+## Usage
+```
+/security scan  # Find issues
+/security fix   # Auto-patch known vulnerabilities
+```

package/.agent/workflows/seo.md

@@ -1,51 +1,28 @@
 ---
-description: Muốn lên Top Google và AI Search? Tối ưu SEO/GEO ngay.
+description: Muốn lên Top Google? Tối ưu SEO ngay.
 ---
 
-# /seo - Strategic Visibility & GEO Optimization
+# /seo - Growth Engine
 
 $ARGUMENTS
 
 ---
 
-## 🟢 PHASE 1: Semantic Discovery
-**Agent**: `seo-specialist` & `explorer-agent`
-**Mission**: Map the "Search Landscape."
-- **Action**: Identify target keywords and "Search Intent."
-- **Action**: Audit current Meta Tags, Headers, and Alt text.
-- **DNA Link**: Check `rules/seo-fundamentals.md`.
+## Task
+Optimize the application for Search Engines (Google, Bing).
 
-## 🟡 PHASE 2: Technical & GEO Audit
-**Agent**: `seo-specialist`
-**Mission**: Optimize for AI Search (GEO).
-- **Checks**:
-  - High-density information?
-  - Cited sources and authority?
-  - JSON-LD / Schema richness?
-  - Core Web Vitals impact?
-
-## 🔵 PHASE 3: Surgical Content Optimization
-**Agent**: `documentation-writer` & `frontend-specialist`
-**Mission**: Implement the findings.
-- **Action**: Update Page Titles, Descriptions, and Semantic HTML.
-- **Action**: Inject structured data (Schema.org).
-
-## 🔴 PHASE 4: Visibility Gating
-**Agent**: `quality-inspector`
-**Mission**: Certify the "Search-Ready" status.
-- **Verification**: Run Lighthouse SEO audit.
-- **Artifact**: Provide a "Search Visibility Report" in the `walkthrough.md`.
-
----
-
-## SEO Objectives:
-- **100/100**: Target Lighthouse SEO score.
-- **GEO-Ready**: Content must be high-signal for AI crawlers.
-- **Accessible**: Semantic HTML is good for both bots and humans.
+### Steps:
+1.  **Sitemap Generation**: Check `sitemap.xml`.
+2.  **Robots.txt**: Verify crawling rules.
+3.  **Meta Tags**: Audit Title/Description on all routes.
+4.  **Schema Markup**: Implement JSON-LD for Products/Articles.
+5.  **Performance**: Check Core Web Vitals (LCP, CLS).
 
 ---
 
-## Examples:
-- `/seo audit product page`
-- `/seo optimize blog post for AI search`
-- `/seo check schema.org implementation`
+## Usage
+```
+/seo audit      # Check current status
+/seo schema     # Generate JSON-LD
+/seo perf       # Optimize Critical Rendering Path
+```

package/.agent/workflows/status.md

@@ -1,47 +1,86 @@
 ---
-description: Dự án đang đến đâu rồi? Xem Dashboard báo cáo chuyên nghiệp.
+description: Dự án đang đến đâu rồi? Xem Dashboard báo cáo.
 ---
 
-# /status - Executive Project Dashboard
+# /status - Show Status
 
 $ARGUMENTS
 
 ---
 
-## 🟢 PHASE 1: Data Aggregation
-**Agent**: `product-owner` & `explorer-agent`
-**Mission**: Gather the "Pulse" of the project.
-- **Action**: Read `task.md`, `walkthrough.md`, and `ERRORS.md`.
-- **Action**: Check git history for recent velocity.
-
-## 🟡 PHASE 2: Logic & Health Assessment
-**Agent**: `product-owner`
-**Mission**: Analyze the "Vitals."
-- **Checks**:
-  - Are milestones being hit?
-  - Is technical debt (ERRORS.md) increasing?
-  - Is the plan still aligned with the goal?
-
-## 🔵 PHASE 3: Dashboard Synthesis
-**Agent**: `product-owner`
-**Mission**: Create a visual overview.
-- **Action**: Generate a Markdown-friendly dashboard with progress bars and status badges.
-
-## 🔴 PHASE 4: Professional Reporting
-**Agent**: `orchestrator`
-**Mission**: Deliver the "Executive Summary."
-- **Artifact**: A concise, professional message to the User summarizing current state and next immediate steps.
+## Task
+
+Show current project and agent status.
+
+### What It Shows
+
+1. **Project Info**
+   - Project name and path
+   - Tech stack
+   - Current features
+
+2. **Agent Status Board**
+   - Which agents are running
+   - Which tasks are completed
+   - Pending work
+
+3. **File Statistics**
+   - Files created count
+   - Files modified count
+
+4. **Preview Status**
+   - Is server running
+   - URL
+   - Health check
 
 ---
 
-## Status Indicators:
-- 🟢 **Healthy**: On track, no blockers.
-- 🟡 **Warning**: Minor delays or increasing errors.
-- 🔴 **Blocked**: Critical dependencies or major bugs.
+## Example Output
+
+```
+=== Project Status ===
+
+📁 Project: my-ecommerce
+📂 Path: C:/projects/my-ecommerce
+🏷️ Type: nextjs-ecommerce
+📊 Status: active
+
+🔧 Tech Stack:
+   Framework: next.js
+   Database: postgresql
+   Auth: clerk
+   Payment: stripe
+
+✅ Features (5):
+   • product-listing
+   • cart
+   • checkout
+   • user-auth
+   • order-history
+
+⏳ Pending (2):
+   • admin-panel
+   • email-notifications
+
+📄 Files: 73 created, 12 modified
+
+=== Agent Status ===
+
+✅ database-architect → Completed
+✅ backend-specialist → Completed
+🔄 frontend-specialist → Dashboard components (60%)
+⏳ test-engineer → Waiting
+
+=== Preview ===
+
+🌐 URL: http://localhost:3000
+💚 Health: OK
+```
 
 ---
 
-## Examples:
-- `/status`
-- `/status focus on infrastructure`
-- `/status show recent errors`
+## Technical
+
+Status uses these scripts:
+- `python .agent/scripts/session_manager.py status`
+- `python .agent/scripts/auto_preview.py status`