A .NET CLI engine that orchestrates multi-step AI pipelines using Claude Code as the execution backend.
Define pipelines in YAML · Compose agents with Markdown bundles · Let Claude handle the rest
- .NET 10 SDK
- Claude Code CLI installed and authenticated (
npm install -g @anthropic-ai/claude-code)
# Clone & build
git clone https://github.com/viamus/code-genesis.git && cd code-genesis
dotnet build
# Run a YAML pipeline
dotnet run --project CodeGenesis.Engine -- run-pipeline examples/hello-world.yml
# Run an ad-hoc task (Plan → Execute → Validate)
dotnet run --project CodeGenesis.Engine -- run "Add retry logic to the HttpClient service"pipeline:
name: "Code Review"
inputs:
task:
description: "What to review"
steps:
- name: "Plan"
prompt: "Create a review plan for: {{task}}"
output_key: "plan"
- name: "Execute"
prompt: "Implement the plan: {{steps.plan}}"
max_turns: 10
output_key: "result"
- name: "Validate"
prompt: "Review: {{steps.result}}"
optional: trueUse
{{variable}}for inputs and{{steps.<key>}}for outputs from previous steps.
Full documentation is available in the Wiki:
| Page | ||
|---|---|---|
| 🚀 | Getting Started | Prerequisites, installation, first pipeline |
| 💻 | CLI Reference | run and run-pipeline commands |
| 📋 | Pipeline YAML Reference | YAML structure, template variables, max_turns |
| 🔀 | Step Types | Simple, Foreach, Parallel, ParallelForeach, Approval |
| 🔌 | MCP Servers | Custom tools via MCP stdio protocol |
| 📦 | Context Bundles | Reusable agent instruction packages |
| ⚙️ | Configuration | Environment variables, appsettings.json, logs |
| 🏗️ | Project Structure | Source tree and architecture |
| 🧪 | Testing | Test project, coverage, CI |
See CONTRIBUTING.md for guidelines on how to contribute.
This project is licensed under the MIT License. See LICENSE for details.