SDD Multi-Agent OpenCode

Spec-Driven Development framework for OpenCode with multi-agent orchestration, DAG-based parallel execution, and free-tier model optimization.

Inspired by spec-kit-command-cursor, rebuilt from scratch for OpenCode's multi-agent architecture.

Layering

This repo is the SDD workflow layer. For base agent behavior, CLI-first scaffolding, status verification, and context-window management, use opencode-agent-rules first.

opencode-agent-rules        # base behavior + context/compaction policy
        ↓
sdd-multiagent-opencode     # SDD commands, agents, skills, templates
        ↓
external UI skills          # impeccable, taste, nothing-design, etc.

What's Different from the Original

Multi-Agent Architecture

Quick Start

npx Install (Recommended)

cd your-project
npx sdd-multiagent-opencode init

This installs both:

• Base OpenCode agent rules (verification, CLI-first, context budget)
• SDD multi-agent workflow (commands, agents, skills, templates)

Flags:

npx sdd-multiagent-opencode init          # Install everything
npx sdd-multiagent-opencode init --sdd-only    # SDD workflow only
npx sdd-multiagent-opencode init --rules-only   # Base rules only
npx sdd-multiagent-opencode init --force         # Overwrite existing
npx sdd-multiagent-opencode init --dry-run       # Preview only

1. Install (Manual)

Option A: Using Commands (Recommended)

If you have OpenCode running, you can use the init commands:

/init-rules   # Install base agent rules first
/init-sdd    # Install SDD workflow layer

Option B: Manual Copy

# Recommended: install base agent rules first
git clone https://github.com/TudeOrangBiasa/opencode-agent-rules.git

# Clone into your project
git clone https://github.com/TudeOrangBiasa/sdd-multiagent-opencode.git

# Or copy the .opencode/ and .sdd/ directories into your project
cp opencode-agent-rules/AGENTS.md /your-project/
cp -r opencode-agent-rules/.opencode/rules /your-project/.opencode/
cp -r opencode-agent-rules/.opencode/plugins /your-project/.opencode/
cp -r sdd-multiagent-opencode/.opencode /your-project/
cp -r sdd-multiagent-opencode/.sdd /your-project/
cp sdd-multiagent-opencode/opencode.json /your-project/

2. Configure

Merge the agent and permission sections from opencode.json into your project's opencode.json.

3. Update Project Profile

Edit .sdd/project-profile.json with your stack:

• framework (e.g., "next", "react", "express")
• language (e.g., "typescript", "javascript")
• packageManager (e.g., "pnpm", "npm", "yarn")
• commands (dev, build, test, lint)

4. Use

# Quick planning (80% of features)
/brief user-auth JWT authentication with login/logout

# Full planning (complex features)
/research database-engine Best database for our use case --deep
/specify user-auth User authentication with login, logout, password reset
/plan user-auth
/tasks user-auth
/implement user-auth

# TDD workflow (test-driven development)
/tasks user-auth --tdd
/implement user-auth --tdd

# Full project roadmap
/sdd-full-plan blog-platform Full-featured blog with CMS and analytics
/execute-parallel blog-platform --until-finish

# Audit
/audit user-auth

Commands

Project Structure

.opencode/
├── opencode.json          # Agent configs, permissions, model assignments
├── agents/
│   ├── sdd-orchestrator.md   # DAG coordinator (GPT 5.5)
│   ├── sdd-planner.md        # Architecture designer (GPT 5.5)
│   ├── sdd-explorer.md       # Codebase investigator (DeepSeek)
│   ├── sdd-implementer.md    # Code generator (DeepSeek)
│   ├── sdd-verifier.md       # Completeness checker (Qwen + Chrome)
│   └── sdd-reviewer.md       # Code reviewer (MiniMax)
├── commands/
│   ├── brief.md              # /brief
│   ├── research.md           # /research
│   ├── specify.md            # /specify
│   ├── plan.md               # /plan
│   ├── tasks.md              # /tasks
│   ├── implement.md          # /implement
│   ├── sdd-full-plan.md      # /sdd-full-plan
│   ├── execute-parallel.md   # /execute-parallel
│   ├── execute-task.md       # /execute-task
│   ├── evolve.md             # /evolve
│   ├── refine.md             # /refine
│   ├── upgrade.md            # /upgrade
│   ├── audit.md              # /audit
│   ├── generate-prd.md       # /generate-prd
│   └── generate-rules.md     # /generate-rules
└── skills/
    ├── sdd-research/SKILL.md
    ├── sdd-planning/SKILL.md
    ├── sdd-implementation/SKILL.md
    ├── sdd-audit/SKILL.md
    └── sdd-evolve/SKILL.md

.sdd/
├── config.json               # Project configuration
├── project-profile.json      # Stack and skill routing config
└── templates/                # Document templates
    ├── feature-brief-v2.md
    ├── spec-compact.md
    ├── plan-compact.md
    ├── tasks-compact.md
    ├── research-compact.md
    ├── todo-compact.md
    ├── roadmap-template.json
    └── roadmap-template.md

specs/
├── active/                   # Features in development
├── backlog/                  # Future features
├── completed/                # Delivered features
└── todo-roadmap/             # Project roadmaps with DAG

Workflows

Quick Planning (80% of features)

/brief → /evolve → /refine → /implement

Full Planning (complex features)

/research → /specify → /plan → /tasks → /implement

Deep Research (unfamiliar domain)

/research --deep → /specify → /plan → /tasks → /implement

Parallel Execution (project roadmap)

/sdd-full-plan → /execute-parallel --until-finish

TDD Workflow (Test-Driven Development)

/brief feature-x → /tasks feature-x --tdd → /implement feature-x --tdd

With --tdd flag:

• /tasks generates TDD-specific tasks with test-first structure
• /implement writes failing test → implements minimum code → verifies pass → refactors

Example:

/tasks user-auth --tdd
/implement user-auth --tdd

Tools Integration

codebase-memory-mcp

Used by sdd-explorer and sdd-implementer for:

• search_graph — find functions, classes, routes by natural language
• trace_path — trace call chains and dependencies
• get_code_snippet — read specific function/class source code
• query_graph — run Cypher queries for complex patterns
• get_architecture — high-level project structure

exa_web_search_exa

Used by sdd-explorer and sdd-research skill for:

• External pattern research
• Documentation lookup
• Technology comparison
• Deep research mode (--deep)

chrome-devtools

Used by sdd-verifier for:

• chrome-devtools_navigate_page — navigate to feature pages
• chrome-devtools_take_screenshot — capture visual state
• chrome-devtools_take_snapshot — analyze DOM structure
• chrome-devtools_list_console_messages — check for errors

License

MIT