docs(ai-autopilot): end-to-end quickstart example (the four layers composed)#105
Merged
Merged
Conversation
…sor + runner + surfaces) A runnable, offline (AiFake) example that composes all four epic layers into one build-a-feature flow: a planner routes subtasks to stack personas, the Supervisor dispatches them, each worker writes Vike/ORM files inside a runner sandbox via runnerTools, and the surfaces render live progress + a background handle. Doubles as the integration contract a real FlueRunner must satisfy. Private workspace example (@gemstack/example-autopilot-quickstart), in turbo typecheck+test. Capstone for epic #97.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Capstone for epic #97. Not a child issue — the proof that the three shipped layers actually compose.
What
A runnable, offline example (
@gemstack/example-autopilot-quickstart) that wires all four ai-autopilot layers into one build-a-feature flow:A lead planner decomposes "Add a paginated Orders page backed by an orders table" and routes each subtask to a stack persona (
universal-orm-modeler,vike-page-builder,ui-intent-designer). The Supervisor dispatches them; each persona worker writes Vike/ORM files inside aFakeRunnersandbox viarunnerTools; progress is rendered throughterminalSink(live) and alaunchAutopilotbackground handle.AiFakescripts the model, so it runs with no API key and is deterministic.Why
It's the integration contract: it exercises
personaInstructions/personaTools+runnerTools+agentPlanner(personaRoster(...))+Supervisor+ surfaces together — so it doubles as the exact end-to-end a realFlueRunnermust satisfy. Going real is a two-line change (dropAiFake, swapFakeRunnerfor a real adapter); nothing else in the flow moves.Verified
pnpm --filter @gemstack/example-autopilot-quickstart startprints the live plan, the files written into the sandbox, a build + preview URL, and the synthesized result. Test suite (2 tests) asserts the layers compose: plan routes by persona, files land in the sandbox, surfaces capture the event stream. Private workspace example, in turbo typecheck + test.Epic status after this
#98 personas, #100 surfaces, #101 teardown, and #99's runner interface are all shipped. The only remaining epic work is the real runner adapters (FlueRunner/WebContainer/Docker) on #99 — infra-gated (needs a live sandbox to verify honestly).