Skip to content

Examples

These are not “let me explain the concept” pages — they are demo repositories you can open and inspect. Each lives under examples/ in the Kaddo repo, with committed .kaddo/ and architecture/ artifacts so you can see exactly what Kaddo produces before running anything.

The four scenarios

ExampleScenarioStateHighlights
Task PilotGreenfield appnewStructured knowledge from day one; full loop
Loyalty LiteExisting apppre-aiscan + agents + Guard drift demo
Old OrdersLegacy MVC applegacyUnderstand-before-change; legacy risks/unknowns
Commerce StackMany reposmultirepomodules map + per-module artifacts

How each example is laid out

  • README.md — the scenario, the commands to run, and what to inspect.
  • expected-flow.md — a command → output → artifact → next step walkthrough.
  • prompt-flow.md — the prompt flow: a Mermaid diagram, the CLI↔LLM split, an input/output table (CLI command / LLM agent / input / output / save as), copy/paste prompt handoffs and the artifact chain — so you can reproduce the loop without guessing which prompt to use.
  • committed sample files (.kaddo/, architecture/, sample/) so the artifacts are visible without running anything.

What is real vs illustrative

  • CLI artifacts (.kaddo/config.yml, architecture/work-items/*.md, the roadmap skeleton, .kaddo/modules.yml, module folders) are exactly what kaddo writes.
  • Agent outputs (capabilities, current-state, filled-in module designs) are illustrative — produced by running the Kaddo agent prompts in your own LLM. Kaddo never calls an LLM, generates code, or understands a system automatically.