opencode-agent-kit 1.0.17 → 1.0.19

package/README.md

@@ -4,7 +4,7 @@
 
 # Agent Kit — Setup Guide
 
-Complete setup guide for the **Agent Kit** — a portable multi-stack AI agent system for OpenCode. Includes 13 specialized agents, 62 skill playbooks, 37 slash commands, and 7 MCP servers.
+Complete setup guide for the **Agent Kit** — a portable multi-stack AI agent system for OpenCode. Includes 13 specialized agents, 63 skill playbooks, 39 slash commands, and 8 MCP servers.
 
 ```bash
 npx opencode-agent-kit init    # One command. Full team.
@@ -191,6 +191,10 @@ After installing `.opencode/`, the following slash commands are available:
 
 # Quality
 /sonarqube-scan [options]             # SonarQube quality scan (issues, security, coverage)
+
+# Memory
+/recall [query]                       # Search past observations and lessons
+/remember [text]                      # Save insight/decision to persistent memory
 ```
 
 ## Using the `.opencode/` Folder
@@ -330,16 +334,17 @@ Skills are stored in `.opencode/skills/` (local in the repo) — no need to sear
 | Agent                      | Key Skills                                                                                                                                                                                                                                     |
 | -------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
 | IT Leader                  | coding-standards, backend-patterns, frontend-patterns                                                                                                                                                                                          |
-| Frontend Developer (Vue)   | coding-standards, frontend-patterns, frontend-design, web-design-guidelines, nuxt-ui, tdd-workflow                                                                                                                                             |
-| Frontend Developer (React) | coding-standards, frontend-patterns, frontend-design, web-design-guidelines, vercel-react-best-practices, vercel-composition-patterns, tdd-workflow                                                                                            |
+| Frontend Developer (Vue)   | coding-standards, frontend-patterns, impeccable, web-design-guidelines, nuxt-ui, tdd-workflow                                                                                                                                                  |
+| Frontend Developer (React) | coding-standards, frontend-patterns, impeccable, web-design-guidelines, vercel-react-best-practices, vercel-composition-patterns, tdd-workflow                                                                                                 |
 | Backend Developer          | coding-standards, backend-patterns, postgres-patterns, security-review                                                                                                                                                                         |
-| UI/UX Designer             | frontend-design, web-design-guidelines, building-components, nuxt-ui                                                                                                                                                                           |
+| UI/UX Designer             | impeccable, web-design-guidelines, building-components, nuxt-ui                                                                                                                                                                                |
 | Code Reviewer / QA         | coding-standards, security-review, tdd-workflow, web-design-guidelines                                                                                                                                                                         |
 | Database Specialist        | postgres-patterns, backend-patterns                                                                                                                                                                                                            |
 | DevOps / Infrastructure    | backend-patterns, coding-standards                                                                                                                                                                                                             |
 | SEO Specialist             | frontend-patterns, web-design-guidelines, nuxt-ui                                                                                                                                                                                              |
 | **Android Developer**      | coding-standards, android-jetpack-compose, edge-to-edge, navigation-3, firebase-basics, play-billing, camera1-to-camerax, r8-analyzer, migrate-xml-views-to-jetpack-compose, gpc-setup, gpc-release-flow, gpc-preflight, gpc-vitals-monitoring |
 | **Flutter Developer**      | coding-standards, flutter (patterns), 10 Flutter skills, 9 Dart skills, firebase-basics                                                                                                                                                        |
+| **All Agents**             | agentmemory (persistent cross-session memory, 53 MCP tools)                                                                                                                                                                                   |
 
 ### Skills Not Required for Core Stack
 
@@ -363,6 +368,7 @@ These can be kept if your team uses multi-stack, but are optional.
 - `verification-loop` — Agent verification cycle
 - `nutrient-document-processing` — Document processing API
 - `project-guidelines-example` — Project guidelines example
+- `agentmemory` — Persistent cross-session memory with 53 MCP tools
 
 ## Skill Locations
 
@@ -388,7 +394,7 @@ If skills are not yet on your machine, copy them from this repo:
 mkdir -p ~/.opencode/skills
 cp -R ./.opencode/skills/coding-standards ~/.opencode/skills/
 cp -R ./.opencode/skills/frontend-patterns ~/.opencode/skills/
-cp -R ./.opencode/skills/frontend-design ~/.opencode/skills/
+cp -R ./.opencode/skills/impeccable ~/.opencode/skills/
 cp -R ./.opencode/skills/web-design-guidelines ~/.opencode/skills/
 cp -R ./.opencode/skills/nuxt-ui ~/.opencode/skills/
 cp -R ./.opencode/skills/security-review ~/.opencode/skills/
@@ -431,8 +437,9 @@ From `.opencode/config.json`, agents use the following MCP servers:
 | `nuxt-ui`    | remote | enabled  | Nuxt UI component docs & examples                    |
 | `playwright` | stdio  | enabled  | Browser automation & E2E testing                     |
 | `postman`    | remote | enabled  | Postman API management (collections, requests, docs) |
-| `figma`      | stdio  | disabled | Figma design file access (optional)                  |
+| `figma`      | remote | disabled | Figma design file access (optional)                  |
 | `stitch`     | remote | disabled | Google Stitch AI design generation (optional)        |
+| `agentmemory`| local  | enabled  | Persistent cross-session memory (53 memory tools)    |
 
 To enable Figma MCP:
 
@@ -440,6 +447,15 @@ To enable Figma MCP:
 export FIGMA_ACCESS_TOKEN="your-token"
 ```
 
+To enable agentmemory:
+
+```bash
+npm install -g @agentmemory/agentmemory
+agentmemory                    # Start the memory server on :3111
+```
+
+Open `http://localhost:3113` for the real-time memory viewer.
+
 To enable Google Stitch MCP:
 
 ```bash

package/bin/commands/init.mjs

@@ -222,12 +222,35 @@ export async function init(options) {
     }
   }
 
+  // 7.6 Install agentmemory globally
+  if (!skipInstall) {
+    console.log(`  🧠 Installing agentmemory (persistent memory)...`);
+    try {
+      execSync(`agentmemory --version`, { stdio: "pipe" });
+      console.log(`     ✓ agentmemory already installed`);
+    } catch {
+      try {
+        execSync(`npm install -g @agentmemory/agentmemory`, {
+          stdio: "pipe",
+          timeout: 60000,
+        });
+        console.log(`     ✓ agentmemory installed globally`);
+      } catch (err) {
+        console.error(`  ⚠  agentmemory global install failed: ${err.message}`);
+        console.error(
+          `    Run "npm install -g @agentmemory/agentmemory" manually`,
+        );
+      }
+    }
+  }
+
   // 8. Update .gitignore
   const gitignorePath = join(targetDir, ".gitignore");
   const gitignoreEntries = [
     ".opencode/*",
     "opencode.json",
     "opencode.example.json",
+    "data/",
   ];
   if (!existsSync(gitignorePath)) {
     writeFileSync(gitignorePath, gitignoreEntries.join("\n") + "\n", "utf-8");
@@ -268,8 +291,12 @@ export async function init(options) {
   console.log(`       • .opencode/commands/  — 35+ slash commands`);
   console.log(`       • .opencode/rules/     — Scoped coding rules`);
   console.log(`       • .opencode/contexts/  — Dev/review/research contexts`);
-  console.log(`       • .opencode/docs/— Agent documentation`);
+  console.log(`       • .opencode/docs/     — Agent documentation`);
+  console.log(`       • .opencode/plugins/  — agentmemory capture plugin (22 hooks)`);
+  console.log(`       • .opencode/hooks/    — agentmemory auto-start wrapper`);
+  console.log(`       • agentmemory (global) — Persistent cross-session memory`);
   console.log(`\n     Next steps:`);
   console.log(`       cd ${targetDir}`);
-  console.log(`       opencode\n`);
+  console.log(`       opencode              # agentmemory auto-starts on first use`);
+  console.log(`       # Viewer: http://localhost:3113`);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "opencode-agent-kit",
-  "version": "1.0.17",
+  "version": "1.0.19",
   "description": "Multi-stack OpenCode agent toolkit — 13 specialized AI agents (Nuxt, React, Node.js, Laravel, CI3, Android, Flutter, DevOps, SEO, SonarQube) with 62 skills, 37 commands, and 7 MCP servers",
   "type": "module",
   "bin": {
@@ -43,6 +43,7 @@
     "devops",
     "seo",
     "sonarqube",
+    "agentmemory",
     "coding-agent",
     "mcp",
     "playwright",

package/template/.opencode/.agents.md

@@ -0,0 +1,55 @@
+## Goal
+- Create a SonarQube quality agent that scans code via SonarQube MCP tools, generates TODOs for issues found, and delegates fixes to relevant subagents
+
+## Constraints & Preferences
+- Must use SonarQube MCP server tools (issues, quality-gates, security-hotspots, duplications, rules, measures, projects, coverage)
+- Must use `todowrite` tool for task tracking (not `todo`)
+- Must follow existing agent structure with Global Rules, Core Identity, What You DO, Delegation Protocol
+- Must integrate with existing agent ecosystem (it-leader.md orchestrator + 13 specialized subagents)
+- Full Scan mode: Issues + Security Hotspots + Duplications + Coverage + Dependencies + Quality Gate
+- Trigger: Manual via `/sonarqube-scan` command or `@sonarqube` mention
+- Landing page (index.html) and all documentation must reflect 13 agents, 37 commands, 7 MCP servers
+
+## Progress
+### Done
+- Fetched SonarQube MCP server documentation and available toolsets
+- Reviewed existing agent configurations (it-leader.md, code-reviewer.md)
+- Created `.opencode/agents/sonarqube-quality.md` (406 lines) — full agent definition with scan modes, delegation matrix, issue lifecycle, output contracts
+- Created `.opencode/commands/sonarqube-scan.md` (65 lines) — slash command with options (--project, --severity, --quick, --rescan)
+- Updated `.opencode/agents/it-leader.md` — added `@sonarqube` to subagents table, capabilities reference, decision tree, verification policy, session subagents list
+- Updated `AGENTS.md` — updated agent count (12→13), added Quality category, SonarQube to subagents list, added SonarQube to MCP servers
+- Updated `index.html` — meta tags, hero badge, subtitle, stats, terminal, orbital nodes, marquee, agent cards, MCP section (all reflect 13 agents, 37 commands, 7 MCP)
+- Updated `README.md` — counts, agent list, custom agents table, commands section
+- Updated `package.json` — description, keywords
+
+### In Progress
+- (none)
+
+### Blocked
+- (none)
+
+## Key Decisions
+- File name: `sonarqube-quality.md` (not `sonarqube-agent.md`) to match naming convention
+- Three operating modes: `quick` (issues only), `full` (all 10 toolsets), `pr` (PR scope)
+- Delegation matrix routes fixes by file extension to 11 domain subagents including `@security-reviewer` for hotspots
+- Issue lifecycle: OPEN → SCANNED → TODO_CREATED → DELEGATED → FIX_APPLIED → RE_SCANNED → VERIFIED/ACCEPTED
+- SonarQube MCP server needs toolsets enabled: `issues,security-hotspots,duplications,coverage,dependency-risks,quality-gates,measures,projects,rules`
+
+## Next Steps
+- (all documentation updated — task complete)
+
+## Critical Context
+- SonarQube MCP tools: `search_sonar_issues_in_projects`, `search_security_hotspots`, `search_duplicated_files`, `get_duplications`, `search_files_by_coverage`, `get_file_coverage_details`, `search_dependency_risks`, `get_quality_gate_status`, `get_component_measures`, `search_my_sonarqube_projects`, `get_rule`, `get_analysis`, `change_sonar_issue_status`, `change_security_hotspot_review_status`
+- Agent must NOT be a coder — only scan, report, create TODOs, and delegate
+- Delegation targets: `@frontend-nuxt`, `@frontend-react`, `@backend`, `@ci3`, `@laravel`, `@android`, `@flutter`, `@database`, `@devops`, `@security-reviewer`
+- Severity mapping: Blocker/Critical → high priority, Major → medium, Minor/Info → low
+- Existing agents count updated from 12 to 13 across all documentation
+
+## Relevant Files
+- `.opencode/agents/sonarqube-quality.md`: New SonarQube quality agent definition (406 lines)
+- `.opencode/commands/sonarqube-scan.md`: New slash command for triggering scans (65 lines)
+- `.opencode/agents/it-leader.md`: Updated orchestrator — added `@sonarqube` to subagents, delegation tree, capabilities
+- `AGENTS.md`: Updated project documentation — agent count, categories, MCP servers
+- `index.html`: Landing page — updated to reflect 13 agents, SonarQube card, 37 commands, 7 MCP servers
+- `README.md`: Setup guide — updated counts, agent list, custom agents table, commands
+- `package.json`: NPM package metadata — updated description and keywords

package/template/.opencode/agents/flutter-developer.md

@@ -90,7 +90,7 @@ If user does not specify mode, infer automatically from task size and risk.
 1. **`coding-standards`** — Universal coding standards and best practices
 2. **`flutter`** — Comprehensive Flutter patterns, architecture, state management, and testing
 3. **`frontend-patterns`** — Mobile UI patterns and component architecture
-4. **`frontend-design`** — Design thinking and aesthetic implementation
+4. **`impeccable`** — Impeccable design intelligence: typography, color, layout, motion, critique, and polish
 5. **`web-design-guidelines`** — UI/UX compliance and accessibility
 
 ### Contextual Skills (Load when needed)

package/template/.opencode/agents/nuxt-frontend-developer.md

@@ -103,7 +103,7 @@ Load these skills by default at the beginning of each session:
 
 1. **`coding-standards`** - Universal coding standards and best practices
 2. **`frontend-patterns`** - Modern Vue/Nuxt patterns and component architecture
-3. **`frontend-design`** - Design thinking and aesthetic implementation
+3. **`impeccable`** - Impeccable design intelligence: typography, color, layout, motion, critique, and polish
 4. **`web-design-guidelines`** - UI/UX compliance and accessibility
 
 For trivial edits, keep skill-loading implicit and avoid verbose announcements.
@@ -124,7 +124,7 @@ Load these skills based on the task context:
 # Session Start Protocol
 
 1. Analyze project structure
-2. Load core skills (coding-standards, frontend-patterns, frontend-design)
+2. Load core skills (coding-standards, frontend-patterns, impeccable)
 3. Connect to MCP servers (Nuxt, Nuxt UI, Playwright)
 4. Identify framework (Nuxt as primary) and load relevant skills
 5. Infer session goals from user request first; ask only when blocked
@@ -1068,7 +1068,7 @@ export function useMarketData(marketId: Ref<string>) {
 
 ## Design & Aesthetics Philosophy
 
-This project uses **Impeccable** (impeccable.style) design intelligence. Impeccable's shared design laws are automatically applied to every UI you build. The `frontend-design` skill is pre-loaded with these principles.
+This project uses **Impeccable** (impeccable.style) design intelligence. Impeccable's shared design laws are automatically applied to every UI you build. The `impeccable` skill is pre-loaded with these principles.
 
 ### Design Thinking Process
 
@@ -1487,7 +1487,7 @@ Before starting work, check if `opencode-agent-kit` has an update:
 Quick context check:
 
 - Project: [Detected framework and version]
-- Loaded skills: coding-standards, frontend-patterns, frontend-design
+- Loaded skills: coding-standards, frontend-patterns, impeccable
 - Ready to: [build components | optimize performance | implement designs | review code]
 
 What are we working on today?

package/template/.opencode/agents/react-frontend-developer.md

@@ -103,7 +103,7 @@ If user does not specify mode, infer automatically from task size and risk.
 
 1. **`coding-standards`** - Universal coding standards and best practices
 2. **`frontend-patterns`** - Modern React/Next.js patterns and component architecture
-3. **`frontend-design`** - Design thinking and aesthetic implementation
+3. **`impeccable`** - Impeccable design intelligence: typography, color, layout, motion, critique, and polish
 4. **`web-design-guidelines`** - UI/UX compliance and accessibility
 
 ### Contextual Skills (Load when needed)
@@ -121,7 +121,7 @@ If user does not specify mode, infer automatically from task size and risk.
 # Session Start Protocol
 
 1. Analyze project structure
-2. Load core skills (coding-standards, frontend-patterns, frontend-design)
+2. Load core skills (coding-standards, frontend-patterns, impeccable)
 3. Identify framework (React/Next.js) and load relevant skills
 4. Infer session goals from user request first; ask only when blocked
 5. Load additional contextual skills as needed
@@ -623,7 +623,7 @@ function LoginForm() {
 
 ## Design & Aesthetics Philosophy
 
-This project uses **Impeccable** (impeccable.style) design intelligence. Impeccable's shared design laws are automatically applied to every UI you build. The `frontend-design` skill is pre-loaded with these principles.
+This project uses **Impeccable** (impeccable.style) design intelligence. Impeccable's shared design laws are automatically applied to every UI you build. The `impeccable` skill is pre-loaded with these principles.
 
 ### Design Thinking Process
 
@@ -958,7 +958,7 @@ React Frontend Developer activated!
 
 Project context:
 - Framework: [React | Next.js | Vite]
-- Loaded skills: coding-standards, frontend-patterns, frontend-design
+- Loaded skills: coding-standards, frontend-patterns, impeccable
 - Ready to: [build components | optimize performance | implement designs | review code]
 
 What are we working on today?

package/template/.opencode/commands/recall.md

@@ -0,0 +1,19 @@
+Search past session observations and lessons for relevant context. Wraps the `memory_smart_search` and `memory_lesson_recall` MCP tools.
+
+## Usage
+
+```
+/recall [query]
+```
+
+## Instructions
+
+1. Call `memory_smart_search` with the query and `limit: 10` (hybrid BM25 + vector + graph search).
+2. Call `memory_lesson_recall` with the same query and `limit: 5` (lesson search).
+3. Combine results and present to the user:
+   - Group by session
+   - Show type, title, and narrative for each observation
+   - Highlight high-importance observations
+   - Show lessons separately with confidence scores
+4. If no results, suggest 2-3 alternative search terms.
+5. **Never hallucinate results.** Only present what the MCP tools actually return.

package/template/.opencode/commands/remember.md

@@ -0,0 +1,19 @@
+Explicitly save an insight, decision, or learning to agentmemory for future sessions. Wraps the `memory_save` MCP tool.
+
+## Usage
+
+```
+/remember [what to remember]
+```
+
+## Instructions
+
+1. Analyze what needs to be remembered — extract the core insight, decision, or fact.
+2. Extract 2-5 searchable concepts (lowercased keyword phrases). Prefer specific terms ("jwt-refresh-rotation" over "auth").
+3. Extract relevant file paths the memory references.
+4. Call `memory_save` with:
+   - `content` — full text to remember (preserve user's phrasing)
+   - `concepts` — extracted concept list
+   - `files` — extracted file list (empty array if none)
+   - `type` — choose from: pattern, preference, architecture, bug, workflow, fact
+5. Confirm the save and show the concepts tagged so the user knows retrieval terms.

package/template/.opencode/docs/frontend/nuxt/COMPLETION_REPORT.md

@@ -64,7 +64,7 @@
 
 - coding-standards
 - frontend-patterns
-- frontend-design
+- impeccable
 - web-design-guidelines
 
 **On-Demand:**

package/template/.opencode/docs/frontend/nuxt/INDEX.md

@@ -28,7 +28,7 @@ Navigasi lengkap untuk semua dokumentasi Frontend Developer Agent.
 .opencode/
 ├── agents/
 │   └── frontend-developer.md  # Agent definition
-└── agent-docs/frontend/nuxt/                # Supporting documentation (this directory)
+└── docs/frontend/nuxt/                # Supporting documentation (this directory)
     ├── INDEX.md               # This file - navigation hub
     ├── README.md              # Complete user guide
     ├── QUICK_START.md         # Get started in 5 minutes

package/template/.opencode/docs/frontend/nuxt/QUICK_START.md

@@ -40,7 +40,7 @@ Agent otomatis akan load skills yang dibutuhkan:
 ```
 ✓ coding-standards
 ✓ frontend-patterns
-✓ frontend-design
+✓ impeccable
 ✓ web-design-guidelines
 ```
 
@@ -179,7 +179,7 @@ Agent punya akses ke 10+ skills dari repository Anda:
 
 - **coding-standards** - Universal best practices
 - **frontend-patterns** - Vue/Nuxt patterns
-- **frontend-design** - Design thinking & aesthetics
+- **impeccable** - Design intelligence (typography, color, layout, motion, critique)
 - **web-design-guidelines** - UI/UX compliance
 
 ### Contextual Skills (On-demand)

package/template/.opencode/docs/frontend/nuxt/README.md

@@ -73,12 +73,12 @@ Agent akan otomatis load skills ini saat session dimulai:
    - Performance optimization
    - Form handling
 
-3. **`frontend-design`**
-   - Design thinking process
-   - Aesthetic guidelines
-   - Typography standards
-   - Animation patterns
-   - Creative implementation
+3. **`impeccable`**
+   - Impeccable design intelligence
+   - Design critique, audit, and polish
+   - Typography and color systems
+   - Layout, motion, and interaction
+   - 23 sub-commands (craft, shape, critique, etc.)
 
 4. **`web-design-guidelines`**
    - UI/UX compliance
@@ -145,7 +145,7 @@ geometric patterns.
 
 Agent akan:
 
-1. Load skill `frontend-design`
+1. Load skill `impeccable`
 2. Choose distinctive fonts (bukan Inter/Roboto)
 3. Create cohesive color palette
 4. Implement purposeful animations
@@ -484,7 +484,7 @@ git commit -m "Update frontend agent: add new skill integration"
 │                                                              │
 │ LOADED SKILLS:                                               │
 │   ✓ coding-standards       ✓ frontend-patterns              │
-│   ✓ frontend-design        ✓ web-design-guidelines          │
+│   ✓ impeccable             ✓ web-design-guidelines          │
 │                                                              │
 │ CONTEXTUAL SKILLS:                                           │
 │   • nuxt-ui                • vercel-composition-patterns    │

package/template/.opencode/docs/frontend/nuxt/SUMMARY.md

@@ -196,7 +196,7 @@ Agent terintegrasi dengan **10+ professional skills**:
 **Core Skills (Auto-loaded):**
 - ✅ coding-standards
 - ✅ frontend-patterns
-- ✅ frontend-design
+- ✅ impeccable
 - ✅ web-design-guidelines
 
 **Contextual Skills (On-demand):**

package/template/.opencode/docs/frontend/nuxt/TESTING_GUIDE.md

@@ -68,7 +68,7 @@ ls ~/.agents/skills/ 2>/dev/null || echo "No additional skills"
 # Expected core skills:
 # - coding-standards
 # - frontend-patterns
-# - frontend-design
+# - impeccable
 # - nuxt-ui
 # - web-design-guidelines
 ```

package/template/.opencode/docs/frontend/nuxt/WORKFLOWS.md

@@ -198,7 +198,7 @@ Identify:
 - Animation opportunities
 - Responsive breakpoints
 
-Load skill `frontend-design` untuk aesthetic analysis.
+Load skill `impeccable` untuk design analysis.
 ```
 
 **Step 2: Setup Structure**

package/template/.opencode/docs/frontend/react/CHEATSHEET.md

@@ -77,7 +77,7 @@
 |-------|-------------|
 | `coding-standards` | Universal coding standards |
 | `frontend-patterns` | React/Next.js patterns |
-| `frontend-design` | Design implementation |
+| `impeccable` | Impeccable design intelligence |
 | `web-design-guidelines` | UI/UX compliance |
 | `shadcn-ui` | shadcn/ui components |
 | `vercel-react-best-practices` | React performance |

package/template/.opencode/docs/frontend/react/INDEX.md

@@ -4,11 +4,11 @@ Navigasi lengkap untuk semua dokumentasi React Frontend Developer Agent (`@front
 
 ## Quick Links
 
-| Kebutuhan | Dokumen |
-|-----------|---------|
-| Baru mulai? | [QUICK_START.md](./QUICK_START.md) |
-| Panduan lengkap? | [README.md](./README.md) |
-| Cari command cepat? | [CHEATSHEET.md](./CHEATSHEET.md) |
+| Kebutuhan               | Dokumen                              |
+| ----------------------- | ------------------------------------ |
+| Baru mulai?             | [QUICK_START.md](./QUICK_START.md)   |
+| Panduan lengkap?        | [README.md](./README.md)             |
+| Cari command cepat?     | [CHEATSHEET.md](./CHEATSHEET.md)     |
 | Data fetching patterns? | [API_PATTERNS.md](./API_PATTERNS.md) |
 
 ## Documentation Structure
@@ -17,7 +17,7 @@ Navigasi lengkap untuk semua dokumentasi React Frontend Developer Agent (`@front
 .opencode/
 ├── agents/
 │   └── react-frontend-developer.md  # Agent definition
-└── agent-docs/frontend/react/       # Supporting documentation (this directory)
+└── docs/frontend/react/       # Supporting documentation (this directory)
     ├── INDEX.md                     # This file — navigation hub
     ├── README.md                    # Complete user guide
     ├── QUICK_START.md               # Get started in 5 minutes
@@ -39,7 +39,7 @@ Navigasi lengkap untuk semua dokumentasi React Frontend Developer Agent (`@front
 
 ## Agent Reference
 
-| Mention | File | Stack |
-|---------|------|-------|
+| Mention           | File                          | Stack                                         |
+| ----------------- | ----------------------------- | --------------------------------------------- |
 | `@frontend-react` | `react-frontend-developer.md` | React 19, Next.js 15, shadcn/ui, Tailwind CSS |
-| `@frontend-nuxt` | `nuxt-frontend-developer.md` | Nuxt 4, Vue 3, Nuxt UI, Tailwind CSS |
+| `@frontend-nuxt`  | `nuxt-frontend-developer.md`  | Nuxt 4, Vue 3, Nuxt UI, Tailwind CSS          |

package/template/.opencode/docs/frontend/react/QUICK_START.md

@@ -15,7 +15,7 @@ Di OpenCode, ketik:
 Agent otomatis akan load skills yang dibutuhkan:
 - `coding-standards`
 - `frontend-patterns`
-- `frontend-design`
+- `impeccable`
 - `web-design-guidelines`
 
 ## Mulai Development!

package/template/.opencode/docs/frontend/react/README.md

@@ -46,7 +46,7 @@ IT Leader akan otomatis mendelegasikan task ke `@frontend-react` ketika project
 Setiap session, agent akan load:
 1. `coding-standards` — Universal coding standards
 2. `frontend-patterns` — React/Next.js patterns
-3. `frontend-design` — Design implementation
+3. `impeccable` — Impeccable design intelligence
 4. `web-design-guidelines` — UI/UX compliance
 
 ### Contextual Skills (load sesuai kebutuhan)

package/template/.opencode/docs/mobile/flutter/README.md

@@ -36,7 +36,7 @@ Cross-platform mobile development dengan Dart, Flutter SDK, dan Material Design
 
 - `coding-standards` — Universal coding standards
 - `frontend-patterns` — Mobile UI patterns
-- `frontend-design` — Design implementation
+- `impeccable` — Impeccable design intelligence
 - `web-design-guidelines` — UI/UX compliance
 - `firebase-basics` — Firebase services
 - `building-components` — Reusable widget libraries

package/template/.opencode/hooks/agentmemory-start.sh

@@ -0,0 +1,17 @@
+#!/usr/bin/env bash
+set -e
+
+AGENTMEMORY_URL="${AGENTMEMORY_URL:-http://localhost:3111}"
+HEALTH_URL="$AGENTMEMORY_URL/agentmemory/livez"
+
+if ! curl -sf "$HEALTH_URL" > /dev/null 2>&1; then
+  npx @agentmemory/agentmemory > /dev/null 2>&1 &
+  for i in $(seq 1 15); do
+    if curl -sf "$HEALTH_URL" > /dev/null 2>&1; then
+      break
+    fi
+    sleep 1
+  done
+fi
+
+exec npx -y @agentmemory/mcp "$@"

package/template/.opencode/instructions/INSTRUCTIONS.md

@@ -380,6 +380,55 @@ flutter analyze                          # Static analysis
 
 ---
 
+## Agentmemory: Persistent Cross-Session Memory
+
+agentmemory provides persistent memory for all agents. It captures session history, saves decisions/insights, and injects relevant context from past sessions into the current session.
+
+### Prerequisites
+
+```bash
+npm install -g @agentmemory/agentmemory   # Install globally
+agentmemory                                # Start server on :3111
+```
+
+### MCP Tools (53 tools)
+
+All agents have access to agentmemory MCP tools prefixed with `agentmemory_memory_`:
+
+| Tool | Purpose |
+|------|---------|
+| `memory_save` | Save insights, decisions, facts to long-term memory |
+| `memory_recall` | Search past observations by keywords |
+| `memory_smart_search` | Hybrid semantic+keyword search for conceptual queries |
+| `memory_sessions` | List recent sessions with status and observation counts |
+| `memory_file_history` | Get past observations about specific files |
+| `memory_lesson_save` | Save a lesson learned with confidence scoring |
+| `memory_lesson_recall` | Search lessons by query, sorted by confidence |
+| `memory_governance_delete` | Delete specific memories (requires confirmation) |
+| `memory_patterns` | Detect recurring patterns across sessions |
+| `memory_consolidate` | Run 4-tier memory consolidation pipeline |
+
+### Available Commands
+
+- `/recall [query]` — Search past observations and lessons
+- `/remember [text]` — Explicitly save an insight to long-term memory
+
+### Auto-Capture Plugin
+
+The `agentmemory-capture.ts` plugin (registered in `opencode.json`) captures 22 lifecycle events automatically:
+- Session lifecycle: created, idle, compacted, updated, deleted, error
+- Messages & prompts: user messages, assistant responses, removed messages
+- Parts & steps: subagent starts, tool calls, reasoning, step-finish, patches, compaction events
+- File enrichment: auto-injects file-specific context into system prompt
+- Permissions: captures permission prompts and replies
+- Tasks & commands: captures todo changes and command execution
+
+### Skills
+
+The `agentmemory` skill (`.opencode/skills/agentmemory/SKILL.md`) teaches agents when and how to use the memory tools effectively.
+
+---
+
 ## opencode-agent-kit Version Check
 
 If `.opencode/.kit-version` exists, your agent toolkit has a recorded installed version.