triples-agentic 2.4.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 PAU Labs
+
+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.

package/README.md

@@ -0,0 +1,326 @@
+# triples-agentic
+
+A platform-agnostic software engineering agent orchestrator, named after the 24-member K-pop group **TripleS**.
+
+11 specialized agents covering the full product-to-delivery lifecycle — from PRD to RFC, task breakdown, implementation across 5 platforms, test cases, and QA.
+
+---
+
+## Agent Roster
+
+| S# | Agent | Persona | Role |
+|----|-------|---------|------|
+| S1 | **SeoYeon** | Engineering Manager | Main Orchestrator |
+| S3 | **JiWoo** | Senior Product Manager | PRD Agent |
+| S5 | **YooYeon** | Staff Engineer / Tech Lead | RFC Agent |
+| S7 | **NaKyoung** | Technical Program Manager | Task Breakdown |
+| S8 | **YuBin** | Principal Frontend Engineer | Frontend Web |
+| S9 | **Kaede** | Principal Backend Engineer | Backend |
+| S12 | **YeonJi** | Senior Android Engineer | Android Native (Kotlin) |
+| S14 | **SoHyun** | Senior iOS Engineer | iOS Native (Swift) |
+| S11 | **Kotone** | Senior Flutter Engineer | Flutter (Dart) |
+| S17 | **Lynn** | QA Lead / Test Lead | Test Cases |
+| S20 | **ShiOn** | Senior QA Automation Engineer | QA Execution |
+
+---
+
+## Workflow
+
+```
+User → SeoYeon (EM)
+         → JiWoo (PRD) [human review loop]
+         → YooYeon (RFC) [human review loop]
+         → NaKyoung (Tasks) [human review loop]
+         → YuBin + Kaede + YeonJi + SoHyun + Kotone [parallel dev]
+           + Lynn (Test Cases) [human review loop]
+         → ShiOn (QA) → Go/No-Go
+```
+
+Full Mermaid diagram: [docs/workflow.md](docs/workflow.md)
+
+---
+
+## Installation
+
+### Install via curl (recommended)
+
+```bash
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash
+```
+
+Checks Node.js ≥ 18, then launches the interactive wizard.
+
+Pass a platform directly to skip the wizard:
+
+```bash
+# Claude Code — project-level
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- claude
+
+# Claude Code — global (all projects on your machine)
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- claude --global
+
+# Cursor AI — global
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- cursor --global
+
+# All platforms at once
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- all
+
+# Update all existing installations (auto-detects platforms and scope)
+curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- update
+```
+
+---
+
+### Interactive wizard
+
+```bash
+npx triples-agentic
+```
+
+```
+╔══════════════════════════════════════════════════╗
+║  TripleS Agentic — Skill Plugin Setup            ║
+║  Software Engineering Agent Orchestrator         ║
+╚══════════════════════════════════════════════════╝
+
+Which coding assistant are you installing for?
+  1. Claude Code
+  2. Cursor AI
+  3. GitHub Copilot
+  4. OpenAI Codex
+  5. Windsurf
+  6. All of the above
+
+Install scope?
+  1. Global — applies to all your projects  (~/.claude/skills/)
+  2. Project — current directory only       (./.claude/skills/)
+```
+
+---
+
+### Direct install (non-interactive)
+
+```bash
+# Project-level (run inside your project directory)
+npx triples-agentic claude      # Claude Code
+npx triples-agentic cursor      # Cursor AI
+npx triples-agentic copilot     # GitHub Copilot
+npx triples-agentic codex       # OpenAI Codex
+npx triples-agentic windsurf    # Windsurf
+npx triples-agentic all         # all platforms
+
+# Global install
+npx triples-agentic claude --global     # → ~/.claude/skills/
+npx triples-agentic cursor --global     # → ~/.cursor/rules/
+npx triples-agentic codex --global      # → ~/.codex/AGENTS.md
+npx triples-agentic windsurf --global   # → ~/.codeium/windsurf/rules/
+
+# Install into a specific directory
+npx triples-agentic claude --target /path/to/project
+
+# Update all existing installations (auto-detects platforms and scope)
+npx triples-agentic update
+```
+
+---
+
+### Install as a global CLI
+
+```bash
+npm install -g triples-agentic
+triples-agentic              # interactive wizard
+triples-agentic claude       # direct install for Claude Code
+```
+
+---
+
+### Files installed per platform
+
+| Platform | Agent skills | Hook config |
+|---|---|---|
+| **Claude Code** | `.claude/skills/*.md` | `.claude/settings.json` (PreToolUse hook) |
+| **Cursor AI** | `.cursor/rules/*.mdc` | `.cursor/rules/triples-safety.mdc` (always-applied rule) |
+| **GitHub Copilot** | `.github/instructions/*.instructions.md` | `.github/instructions/triples-safety.instructions.md` |
+| **OpenAI Codex** | `AGENTS.md` (project) / `~/.codex/AGENTS.md` (global) | `.codex/config.toml` / `~/.codex/config.toml` (PreToolUse hook) |
+| **Windsurf** | `.windsurfrules` | `.windsurf/hooks.json` (pre_run_command hook) |
+
+Global install paths:
+
+| Platform | Skills | Hook config |
+|---|---|---|
+| Claude Code | `~/.claude/skills/` | `~/.claude/settings.json` |
+| Cursor AI | `~/.cursor/rules/` | `~/.cursor/rules/triples-safety.mdc` |
+| OpenAI Codex | `~/.codex/AGENTS.md` | `~/.codex/config.toml` |
+| Windsurf | `~/.codeium/windsurf/rules/` | `~/.codeium/windsurf/hooks.json` |
+
+---
+
+## Safety Guardrails
+
+Every install includes safety enforcement that blocks dangerous shell commands before they run.
+
+### How it works per platform
+
+| Platform | Mechanism | Enforcement |
+|---|---|---|
+| **Claude Code** | `PreToolUse` hook in `.claude/settings.json` | Hard — harness blocks the command |
+| **OpenAI Codex** | `PreToolUse` hook in `.codex/config.toml` | Hard — same engine as Claude Code |
+| **Windsurf** | `pre_run_command` hook in `.windsurf/hooks.json` | Hard — exit code 2 blocks execution |
+| **Cursor AI** | `alwaysApply: true` rule in `.cursor/rules/` | Soft — AI-level instruction |
+| **GitHub Copilot** | `applyTo: "**"` instruction in `.github/instructions/` | Soft — AI-level instruction |
+
+### Blocked commands
+
+- **Filesystem** — `rm -rf`, `rm -fr`, recursive force delete
+- **Git destructive** — `git reset --hard`, `git checkout --`, `git restore .`
+- **Git push force** — `git push --force`, `git push -f`
+- **SQL destructive** — `DROP TABLE`, `DROP DATABASE`, `DELETE FROM <table>;`
+- **Store publish** — `fastlane deliver/supply`, `gradlew publishBundle`, `xcrun altool`
+- **Package publish** — `npm publish`, `npx ... publish`
+
+Safe commands (`npm test`, `git push origin branch`, `flutter test`, etc.) pass through without interruption.
+
+### Adding or editing safety rules
+
+All hook definitions live in [`src/hooks/`](src/hooks/):
+
+- `dangerous-commands.json` — per-platform hook configs (Claude Code, Codex, Windsurf)
+- `dangerous-commands.md` — text-based safety rules (Cursor, Copilot)
+
+Edit these files and reinstall to update the rules across all platforms.
+
+---
+
+## Usage
+
+### Full pipeline (Claude Code)
+```
+/seoyeon run
+```
+SeoYeon walks you through the complete workflow and delegates to each agent.
+
+### Individual agents
+```
+/jiwoo-prd       Create, review, and finalize a PRD
+/yooyeon-rfc     Create, review, and finalize an RFC
+/nakyoung-tasks  Create task breakdown with story points and estimates
+/yubin-frontend  Implement frontend web features
+/kaede-backend   Implement backend API and services
+/yeonji-android  Implement Android (Kotlin + Compose) features
+/sohyun-ios      Implement iOS (Swift + SwiftUI) features
+/kotone-flutter  Implement Flutter (Dart) cross-platform features
+/lynn-testcase   Create, review, and finalize test cases
+/shion-qa        Execute tests and produce Go/No-Go report
+/seoyeon status  Check current run state
+```
+
+### With other coding assistants
+Ask for the agent by name — e.g., "Act as JiWoo and create a PRD for [description]"
+
+---
+
+## Human-in-the-Loop Gates
+
+JiWoo, YooYeon, NaKyoung, and Lynn all have built-in review loops:
+
+1. Agent creates artifact
+2. Agent reviews against quality gate checklist
+3. Agent evaluates: `READY` or `GAPS: [list]`
+4. On gaps: agent presents numbered list to user with specific questions
+5. User provides clarifications
+6. Agent updates and loops → repeat until `READY`
+
+This ensures PRD, RFC, task breakdown, and test cases are implementation-ready before moving forward.
+
+---
+
+## Project Structure
+
+```
+triples-agentic/
+├── install.sh                 # curl installer
+├── src/
+│   ├── agents/                # 11 agent definitions (identity, persona, knowledge, tools, workflow)
+│   │   ├── seoyeon.md
+│   │   ├── jiwoo-prd.md
+│   │   ├── yooyeon-rfc.md
+│   │   ├── nakyoung-tasks.md
+│   │   ├── yubin-frontend.md
+│   │   ├── kaede-backend.md
+│   │   ├── yeonji-android.md
+│   │   ├── sohyun-ios.md
+│   │   ├── kotone-flutter.md
+│   │   ├── lynn-testcase.md
+│   │   └── shion-qa.md
+│   ├── hooks/                 # Safety guardrail definitions (source of truth)
+│   │   ├── dangerous-commands.json   # per-platform hook configs
+│   │   └── dangerous-commands.md     # text-based safety rules
+│   ├── knowledge/             # Domain expertise loaded by agents (knowledge-only .md files)
+│   │   ├── general/           # coding principles shared by all developer agents
+│   │   │   └── dry.md, kiss.md, yagni.md, solid.md, slap.md, tdd.md, …
+│   │   ├── planning/          # SeoYeon, JiWoo, YooYeon, NaKyoung
+│   │   │   └── orchestration.md, prd-writing.md, rfc-writing.md, task-decomposition.md, …
+│   │   ├── web/
+│   │   │   ├── frontend/      # YuBin
+│   │   │   │   └── frontend-components.md, frontend-state.md, web-accessibility.md, …
+│   │   │   └── backend/       # Kaede
+│   │   │       └── backend-structure.md, api-design.md, api-security.md, …
+│   │   ├── mobile/
+│   │   │   ├── android/       # YeonJi
+│   │   │   │   └── android-architecture.md, kotlin-core.md, kotlin-concurrency.md, …
+│   │   │   ├── ios/           # SoHyun
+│   │   │   │   └── ios-architecture.md, swift-core.md, swift-concurrency.md, …
+│   │   │   └── flutter/       # Kotone
+│   │   │       └── flutter-architecture.md, dart-core.md, dart-async.md, …
+│   │   └── quality/           # Lynn, ShiOn
+│   │       └── testing-strategy.md, test-case-writing.md, qa-execution.md, …
+│   ├── templates/             # Output document templates
+│   │   ├── prd.md             # → workspace/PRD.md
+│   │   ├── rfc.md             # → workspace/RFC.md
+│   │   ├── task-breakdown.md  # → workspace/TASK_BREAKDOWN.md
+│   │   └── test-case.md       # → workspace/TEST_CASES.md
+│   └── bin/
+│       └── setup.js           # Platform installer CLI
+├── docs/
+│   └── workflow.md            # Full workflow diagram + agent roster
+└── CHANGELOG.md
+
+── Generated by installer (not committed to your repo) ────────────────────
+your-project/
+├── .claude/
+│   ├── skills/                # Claude Code agent + knowledge skills
+│   └── settings.json          # Claude Code PreToolUse safety hook
+├── .cursor/rules/             # Cursor AI agent rules + safety rule
+├── .github/instructions/      # Copilot agent instructions + safety instruction
+├── .codex/config.toml         # Codex PreToolUse safety hook
+├── .windsurf/hooks.json       # Windsurf pre_run_command safety hook
+├── AGENTS.md                  # Codex agent context
+└── .windsurfrules             # Windsurf agent rules
+```
+
+---
+
+### `agents/` — Behavioral definitions
+
+Each agent file defines: identity, persona, knowledge references, tool guardrails, skills (workflows), and handoff signals. Domain content lives in `knowledge/` — agents only reference it.
+
+Each agent includes a `## Tools` section that specifies which tools to use and which to avoid (e.g., planning agents never use `Bash`; QA agent never edits source files).
+
+### `hooks/` — Safety guardrails (source of truth)
+
+Hook definitions live in `src/hooks/` as platform-agnostic source files. The installer reads them and generates the appropriate config for each platform:
+
+- **`.json` files** carry per-platform hook configs under a `platforms` key (`claude`, `codex`, `windsurf`)
+- **`.md` files** carry text-based safety rules for platforms without executable hooks (`cursor`, `copilot`)
+
+### `knowledge/` — Domain expertise
+
+Pure reference `.md` files — no trigger logic, no skills framework. Agents declare which files to load via `<!-- knowledge: ... -->` metadata in their agent file. The knowledge files travel with the agents during install.
+
+---
+
+## Inspired by
+
+[TripleS](https://triples.fandom.com/wiki/TripleS) — a 24-member multi-national K-pop group under MODHAUS, operating on a decentralized system where members rotate between subunits based on fan participation.
+
+Just like TripleS subunits activate for specific projects, TripleS agents activate for specific phases of software delivery.

package/docs/workflow.md

@@ -0,0 +1,163 @@
+# TripleS Agent Orchestration Workflow
+
+## Agent Roster
+
+| S# | Agent | Persona | Role | Slash Command |
+|----|-------|---------|------|---------------|
+| S1 | **SeoYeon** | Engineering Manager | Main Orchestrator | `/seoyeon` |
+| S3 | **JiWoo** | Senior Product Manager | PRD Agent | `/jiwoo-prd` |
+| S5 | **YooYeon** | Staff Engineer / Tech Lead | RFC Agent | `/yooyeon-rfc` |
+| S7 | **NaKyoung** | Technical Program Manager | Task Breakdown | `/nakyoung-tasks` |
+| S8 | **YuBin** | Principal Frontend Engineer | Frontend Web Dev | `/yubin-frontend` |
+| S9 | **Kaede** | Principal Backend Engineer | Backend Dev | `/kaede-backend` |
+| S12 | **YeonJi** | Senior Android Engineer | Android Native | `/yeonji-android` |
+| S14 | **SoHyun** | Senior iOS Engineer | iOS Native | `/sohyun-ios` |
+| S11 | **Kotone** | Senior Flutter Engineer | Flutter Dev | `/kotone-flutter` |
+| S17 | **Lynn** | QA Lead / Test Lead | Test Case Agent | `/lynn-testcase` |
+| S20 | **ShiOn** | Senior QA Automation Engineer | QA Execution | `/shion-qa` |
+
+---
+
+## Full Orchestration Workflow
+
+```mermaid
+flowchart TD
+    User([👤 User Input]) --> SeoYeon
+
+    SeoYeon["🎯 SeoYeon\nEngineering Manager\n/seoyeon"]
+
+    SeoYeon -->|"Delegate: Create PRD"| JiWoo
+
+    subgraph PRD_LOOP["📋 PRD Phase — JiWoo (Senior PM)"]
+        JiWoo["JiWoo\n/jiwoo-prd"]
+        JiWoo --> PRDCreate["1. Create PRD\n(from prd_template.md)"]
+        PRDCreate --> PRDReview["2. Review PRD\n(quality checklist)"]
+        PRDReview --> PRDEval{3. Evaluate:\nAll quality gates\npassed?}
+        PRDEval -->|"❌ GAPS FOUND"| HumanPRD["🧑 Human Review\nJiWoo presents gap list\nUser adds missing context"]
+        HumanPRD -->|"4. Update PRD"| PRDReview
+    end
+
+    PRDEval -->|"✅ READY"| YooYeon
+
+    subgraph RFC_LOOP["⚙️ RFC Phase — YooYeon (Staff Engineer)"]
+        YooYeon["YooYeon\n/yooyeon-rfc"]
+        YooYeon --> RFCCreate["1. Create RFC\n(from rfc_template.md)"]
+        RFCCreate --> RFCReview["2. Review RFC\n(quality checklist)"]
+        RFCReview --> RFCEval{3. Evaluate:\nAll quality gates\npassed?}
+        RFCEval -->|"❌ GAPS FOUND"| HumanRFC["🧑 Human Review\nYooYeon presents\ntechnical gaps\nUser decides"]
+        HumanRFC -->|"4. Update RFC"| RFCReview
+    end
+
+    RFCEval -->|"✅ READY"| NaKyoung
+
+    subgraph TASK_LOOP["📊 Task Breakdown Phase — NaKyoung (TPM)"]
+        NaKyoung["NaKyoung\n/nakyoung-tasks"]
+        NaKyoung --> TaskCreate["1. Create Tasks\n(story points + estimates)"]
+        TaskCreate --> TaskReview["2. Review Tasks\n(readiness checklist)"]
+        TaskReview --> TaskEval{3. Evaluate:\nAll tasks clear\n& estimable?}
+        TaskEval -->|"❌ GAPS FOUND"| HumanTask["🧑 Human Review\nNaKyoung presents\nambiguous tasks\nUser clarifies"]
+        HumanTask -->|"4. Update Tasks"| TaskReview
+    end
+
+    TaskEval -->|"✅ READY"| DevPhase
+    TaskEval -->|"✅ READY"| TestPhase
+
+    subgraph DevPhase["⚡ Development Phase — Parallel"]
+        YuBin["🌐 YuBin\nPrincipal Frontend\n/yubin-frontend\n\nReact/Vue/Angular\nTailwind, TypeScript"]
+        Kaede["🖥️ Kaede\nPrincipal Backend\n/kaede-backend\n\nNode/Python/Go\nPostgreSQL, REST"]
+        YeonJi["🤖 YeonJi\nSenior Android\n/yeonji-android\n\nKotlin + Jetpack Compose\nMaterial Design 3"]
+        SoHyun["🍎 SoHyun\nSenior iOS\n/sohyun-ios\n\nSwift + SwiftUI\nApple HIG"]
+        Kotone["🦋 Kotone\nSenior Flutter\n/kotone-flutter\n\nDart + Flutter\nAndroid + iOS + Web"]
+    end
+
+    subgraph TestPhase["📝 Test Case Phase — Lynn (QA Lead)"]
+        Lynn["Lynn\n/lynn-testcase"]
+        Lynn --> TCCreate["1. Create Test Cases\n(from test_case_template.md)"]
+        TCCreate --> TCReview["2. Review Test Cases\n(quality checklist)"]
+        TCReview --> TCEval{3. Evaluate:\nAll PRD criteria\ncovered?}
+        TCEval -->|"❌ GAPS FOUND"| HumanTC["🧑 Human Review\nLynn presents\nmissing scenarios\nUser adds context"]
+        HumanTC -->|"4. Update Test Cases"| TCReview
+    end
+
+    TCEval -->|"✅ READY"| ShiOn
+    YuBin --> ShiOn
+    Kaede --> ShiOn
+    YeonJi --> ShiOn
+    SoHyun --> ShiOn
+    Kotone --> ShiOn
+
+    subgraph QA["🔍 QA Phase — ShiOn (Senior QA Automation)"]
+        ShiOn["ShiOn\n/shion-qa"]
+        ShiOn --> Smoke["Smoke Tests\n(P0 only)"]
+        Smoke --> SmokeGate{P0 pass?}
+        SmokeGate -->|"❌ Fail"| BlockRelease["🛑 Block Release\nReport P0 failures\nto SeoYeon"]
+        SmokeGate -->|"✅ Pass"| FullRun["Full Test Suite\nP1 → P2 → P3\n+ Exploratory"]
+        FullRun --> GoNoGo{Go/No-Go\nAssessment}
+        GoNoGo -->|"✅ GO"| Done
+        GoNoGo -->|"❌ NO-GO"| BlockRelease2["🛑 No-Go Report\nOpen defects listed\nFix & re-test"]
+    end
+
+    Done(["✅ Delivery Complete\nQA_REPORT.md\nDELIVERY_SUMMARY.md"])
+```
+
+---
+
+## Human-in-the-Loop Gates
+
+Human review is required at four stages. Each gate follows the same pattern:
+
+1. Agent **creates** artifact using its template
+2. Agent **reviews** against its quality gate checklist
+3. Agent **evaluates**: all gates pass → `READY`; any fail → `GAPS: [numbered list]`
+4. Agent **presents** gaps to the human with specific questions
+5. Human **provides** clarifications
+6. Agent **updates** artifact and loops back to step 2
+7. Loop exits when `READY`
+
+| Gate | Agent | Artifact |
+|------|-------|---------|
+| PRD Review | JiWoo (Senior PM) | `workspace/PRD.md` |
+| RFC Review | YooYeon (Staff Engineer) | `workspace/RFC.md` |
+| Task Breakdown Review | NaKyoung (TPM) | `workspace/TASK_BREAKDOWN.md` |
+| Test Case Review | Lynn (QA Lead) | `workspace/TEST_CASES.md` |
+
+---
+
+## Workspace Artifacts
+
+```
+workspace/
+├── PRD.md                    ← JiWoo
+├── RFC.md                    ← YooYeon
+├── TASK_BREAKDOWN.md         ← NaKyoung
+├── TEST_CASES.md             ← Lynn
+├── BUGS/
+│   └── BUG-[ID].md          ← ShiOn (one per defect)
+├── QA_REPORT.md              ← ShiOn
+└── DELIVERY_SUMMARY.md       ← SeoYeon
+```
+
+---
+
+## Quick Start
+
+### Full pipeline
+```
+/seoyeon run
+```
+SeoYeon walks you through the entire workflow, delegating to each agent in sequence.
+
+### Individual agents
+```
+/jiwoo-prd       → Start or resume PRD creation
+/yooyeon-rfc     → Start or resume RFC from PRD
+/nakyoung-tasks  → Start or resume task breakdown
+/yubin-frontend  → Implement frontend web tasks
+/kaede-backend   → Implement backend tasks
+/yeonji-android  → Implement Android tasks
+/sohyun-ios      → Implement iOS tasks
+/kotone-flutter  → Implement Flutter tasks
+/lynn-testcase   → Start or resume test case creation
+/shion-qa        → Execute QA against test cases + dev output
+/seoyeon status  → Check current run state
+```

package/install.sh

@@ -0,0 +1,98 @@
+#!/usr/bin/env bash
+# TripleS Agentic — Installer
+#
+# Usage:
+#   curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash
+#   curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- claude
+#   curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- claude --global
+#   curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- all
+#   curl -fsSL https://raw.githubusercontent.com/pauplayground007/triples-agentic/main/install.sh | bash -s -- update
+
+set -euo pipefail
+
+# ── Terminal colours (disabled if no TTY or TERM is dumb) ────────────────────
+if [ -t 1 ] && [ "${TERM:-}" != "dumb" ]; then
+  RED='\033[0;31m'; GREEN='\033[0;32m'; YELLOW='\033[1;33m'
+  CYAN='\033[0;36m'; BOLD='\033[1m'; NC='\033[0m'
+else
+  RED=''; GREEN=''; YELLOW=''; CYAN=''; BOLD=''; NC=''
+fi
+
+MIN_NODE=18
+
+# ── Helpers ──────────────────────────────────────────────────────────────────
+err()     { printf "${RED}✖ %s${NC}\n" "$*" >&2; exit 1; }
+warn()    { printf "${YELLOW}⚠ %s${NC}\n" "$*"; }
+info()    { printf "${CYAN}→${NC} %s\n" "$*"; }
+success() { printf "${GREEN}✔${NC} %s\n" "$*"; }
+
+# ── Banner ────────────────────────────────────────────────────────────────────
+banner() {
+  printf "\n${BOLD}╔══════════════════════════════════════════════════╗${NC}\n"
+  printf "${BOLD}║  TripleS Agentic — Install                       ║${NC}\n"
+  printf "${BOLD}║  Software Engineering Agent Orchestrator         ║${NC}\n"
+  printf "${BOLD}╚══════════════════════════════════════════════════╝${NC}\n\n"
+}
+
+# ── Prerequisite checks ───────────────────────────────────────────────────────
+check_node() {
+  if ! command -v node &>/dev/null; then
+    err "Node.js is required but not installed.
+
+  Install it from : https://nodejs.org
+  macOS (Homebrew): brew install node
+  Linux (nvm)     : curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.0/install.sh | bash"
+  fi
+
+  local major
+  major=$(node -e "process.stdout.write(String(process.versions.node.split('.')[0]))")
+  if [ "$major" -lt "$MIN_NODE" ]; then
+    err "Node.js v${MIN_NODE}+ required (found v${major}).
+  Update from: https://nodejs.org"
+  fi
+
+  success "Node.js $(node --version)"
+}
+
+check_npx() {
+  if ! command -v npx &>/dev/null; then
+    err "npx not found. Update npm:  npm install -g npm@latest"
+  fi
+  success "npx $(npx --version)"
+}
+
+check_platform() {
+  case "$(uname -s)" in
+    Darwin|Linux) ;;
+    MINGW*|MSYS*|CYGWIN*)
+      warn "Windows detected — running via Git Bash / MSYS2."
+      warn "For full support use WSL2 or run:  npx triples-agentic"
+      ;;
+    *)
+      warn "Unknown platform — proceeding anyway."
+      ;;
+  esac
+}
+
+# ── Main ──────────────────────────────────────────────────────────────────────
+main() {
+  banner
+  check_platform
+  check_node
+  check_npx
+
+  printf "\n"
+  info "Launching triples-agentic setup…"
+  printf "\n"
+
+  # Forward every argument straight to the npm package:
+  #   (no args)          → interactive wizard
+  #   claude             → Claude Code, project-level
+  #   claude --global    → Claude Code, global
+  #   cursor --global    → Cursor AI, global
+  #   all                → all platforms, project-level
+  #   update             → update all existing installations
+  npx --yes github:pauplayground007/triples-agentic "$@"
+}
+
+main "$@"

package/package.json

@@ -0,0 +1,54 @@
+{
+  "name": "triples-agentic",
+  "version": "2.4.0",
+  "description": "TripleS software engineering agent orchestrator — skill plugin for Claude Code, Cursor, Copilot, Codex, and Windsurf",
+  "type": "module",
+  "bin": {
+    "triples-agentic": "./src/bin/setup.js"
+  },
+  "files": [
+    "src/",
+    "docs/",
+    "README.md",
+    "install.sh"
+  ],
+  "scripts": {
+    "setup": "node src/bin/setup.js",
+    "setup:claude": "node src/bin/setup.js claude",
+    "setup:claude:global": "node src/bin/setup.js claude --global",
+    "setup:cursor": "node src/bin/setup.js cursor",
+    "setup:cursor:global": "node src/bin/setup.js cursor --global",
+    "setup:copilot": "node src/bin/setup.js copilot",
+    "setup:codex": "node src/bin/setup.js codex",
+    "setup:windsurf": "node src/bin/setup.js windsurf",
+    "setup:windsurf:global": "node src/bin/setup.js windsurf --global",
+    "setup:all": "node src/bin/setup.js all"
+  },
+  "keywords": [
+    "triples-agentic",
+    "software-engineering",
+    "agent",
+    "orchestrator",
+    "skill",
+    "plugin",
+    "claude",
+    "copilot",
+    "cursor",
+    "codex",
+    "windsurf",
+    "prd",
+    "rfc"
+  ],
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/pauplayground007/triples-agentic.git"
+  },
+  "bugs": {
+    "url": "https://github.com/pauplayground007/triples-agentic/issues"
+  },
+  "homepage": "https://github.com/pauplayground007/triples-agentic#readme"
+}