OmniLearn Workflow

AI-powered adaptive learning workflows for OpenCode. Uses multi-agent orchestration to create personalized learning roadmaps, generate hands-on assignments, and track progress across any skill.

GitHub: https://github.com/BlackPool25/OmniLearn

Quick Start

# Run the installer — it will set up everything interactively
npx omnilearn-workflow

# The installer will:
#   ✓ Check for OpenCode (and install it if missing)
#   ✓ Install all /omnilearn-* commands
#   ✓ Configure your learning directory
#   ✓ Detect optional dependencies

# Then open OpenCode and create your first roadmap:
opencode
/omnilearn-roadmap I want to learn Rust

One-liner (auto-install with defaults)

npx omnilearn-workflow --yes

Available Commands

| Command | Purpose | |---------|---------| | /omnilearn-init | Set up your learning directory and config | | /omnilearn-roadmap | Create a personalized, research-backed learning roadmap | | /omnilearn-roadmap-edit | Edit an existing roadmap without losing progress | | /omnilearn-start | Interactive learning: hands-on assignments, Q&A, progress tracking | | /omnilearn-refine | Deep-dive questions and subtopic refinements |

How It Works

OmniLearn uses multi-agent orchestration (Sisyphus → subagents) to:

1. Research — Parallel subagents use web search + Context7 to find everything needed for real-world proficiency
2. Adapt — Scans your existing skills, reads your preferences, and skips what you already know
3. Integrate — Combines skills you already have with new ones (e.g., Python + FastAPI = full-stack ML API)
4. Generate — Creates hands-on assignments at 3 difficulty levels: Basic → Intermediate → Real-World
5. Track — Every assignment completion and session is logged in topic-progress.md

Architecture

.omnilearn/                          ← In your configured learning directory
├── UserPreferences.md               ← Global preferences (auto-learned)
└── <skill>/
    ├── roadmap.md                   ← Master roadmap
    ├── SkillPreferences.md          ← Per-skill preferences
    ├── progress-index.md            ← Overview index
    ├── runs/                        ← Action logs only
    └── topics/<topic>/
        ├── topic-roadmap.md         ← Detailed subtopic plan
        ├── topic-progress.md        ← 🔑 Progress lives here
        ├── assignments/             ← Hands-on tasks with tests
        └── runs/                    ← Session action logs

Requirements

• OpenCode (installer can auto-install this for you)
• Oh-My-OpenAgent (recommended)
• Node.js >= 18

Configuration

OmniLearn stores its config at ~/.config/opencode/omnilearn.json:

{
  "learningDirectory": "/absolute/path/to/learning",
  "setupDate": "2026-07-04",
  "version": "1"
}

The installer can configure this for you during setup, or you can run /omnilearn-init in OpenCode.

CLI Flags

| Flag | Description | |------|-------------| | --yes, -y | Auto-install with defaults (non-interactive) | | --help, -h | Show usage help | | --version, -v | Show package version | | --check, -c | Verify install health and detect issues |

License

MIT