feat(wizard): shared wizard shell, used by project create and add runtime - #2284
Merged
Merged
Conversation
agentcore-devx-automation
Bot
added
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
![agentcore-devx-automation[bot]](https://avatars.githubusercontent.com/u/2232217?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
AgentCore Harness Review
Verdict: Looks good
Nice refactor. The shared shell is small and legible, the shape it enforces on callers (one field per step, fields own their useInput) is well-documented, and the extraction of toAddRuntimeInput keeps the flag path and wizard path validating through the same code — which is the interesting correctness win here. Position-by-key handling of conditional branches is a nice touch, and it's covered by wizard.test.tsx.
A few things I checked and am happy with:
- Tests use real dependencies.
runtime.screen.test.tsxwrites into a real project viacreateGatewayProjectTestHarnessand asserts the on-disk spec + scaffolded files;wizard.test.tsxdrives a syntheticTestWizardend-to-end rather than mocking the shell. No I/O mocks beyond what the harness already provides. - Double-submit guard is sound.
submittingref is reset infinally, so a retry after an error path still works. ExitOnErrorroutes failures throughexit(error)so they surface via the normal CLI error path instead of a React stack trace — matches how the existing wizards behaved.project createbehaviour is preserved. The old screen test still passes as-is per the PR description; theModelFieldcompound step is kept intentionally, which is the right call given the shell has no focus model.- Telemetry. No new instrumentation, but the existing project handlers aren't instrumented either, so this isn't a regression.
Non-blocking observations (won't ask for changes):
useMemoaroundstepElements/stepswon't actually memoize becausechildrenis a fresh array literal each render. Harmless; the work is trivial.- Success footer label went from
[enter] exit(old create) to[enter] continue(new shared default). Minor copy change; fine either way.
Nothing that needs to change before merging.
agentcore-devx-automation
Bot
removed
the
agentcore-harness-reviewing
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## refactor #2284 +/- ##
============================================
+ Coverage 97.06% 97.08% +0.01%
============================================
Files 569 576 +7
Lines 39322 39519 +197
============================================
+ Hits 38167 38366 +199
+ Misses 1155 1153 -2 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
notgitika
commented
Sep 11, 2026
| // rather than inside renderTui so that a piped or CI run stays headless and | ||
| // reports a missing required flag as the usage error it is, instead of | ||
| // renderTui's "interactive mode requires a TTY". | ||
| export function withTuiWhenInteractive(core: Core, io: AppIO): Middleware { |
Contributor
Author
There was a problem hiding this comment.
withTuiOnEmptyFlagsAndArgs above opens the TUI even without a TTY, so a piped/CI add runtime would fail with "requires a TTY" instead of the missing --name error. This is that middleware behind a TTY check. create and status each had it hand-rolled; now they share it.
notgitika
marked this pull request as ready for review
September 11, 2026 19:06
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
tejaskash
reviewed
Sep 11, 2026
jariy17
previously approved these changes
Sep 14, 2026
jariy17
left a comment
jariy17
left a comment
Contributor
There was a problem hiding this comment.
Really good left some follow up nits
| project.handler(createAddProjectResourceHandler(config)); | ||
| // create wizard; any user-supplied flag, --json, or a non-TTY invocation keeps | ||
| // the headless handler (see withTuiWhenInteractive). | ||
| const tuiWhenInteractive = withTuiWhenInteractive(core, io); |
Contributor
There was a problem hiding this comment.
I thought the middleware handles when jo flags are passed. Do we still need this?
Contributor
Author
There was a problem hiding this comment.
You’re right that the existing middleware handles the no-flags check. This wrapper only adds the TTY gate: interactive invocations open the TUI, while non-TTY invocations use the normal headless handler.
If this is being centralized elsewhere, we can remove it here. I can make this change in a follow up PR by adding it into the existing middleware. What do you think?
…time Adds src/components/wizard: a Wizard shell that owns the step list, position, key handling and the form → running → success | error phases, with Step and the TextField/ChoiceField/Summary fields. Fields publish their own key hints and own their step's input handling, so a screen only declares questions. Migrates project create onto it (−380 lines of its own controller) and adds the first add wizard, project add runtime: name → template → description → review. Both paths build their input through one builder — create through resolveScaffoldHarnessInput/resolveRuntimeTemplateShortcut as before, add runtime through the new toAddRuntimeInput — so a value the flag path accepts is not one a wizard rejects. The runtime wizard scaffolds from a template; the JSON infrastructure flags, the model provider/api-key and the Bedrock Agent import stay on the flag path. Also: KeyValueTable grows into the space it is given, so its percentage key cap resolves when it is laid out as a row item, and FormTextInput/ FormRadioGroup render bare when given no name or help text, since the step already asks the question.
…lidation - a bare `project add runtime` on a TTY now opens the wizard: the add router uses withTuiWhenInteractive, which is withTuiOnEmptyFlagsAndArgs behind a TTY gate. Flags, --json and non-TTY runs stay headless. The same middleware replaces the hand-rolled dispatch shims on create and status. - text fields validate the value as typed rather than a trimmed copy, so a step cannot accept " my_agent " and then submit it. - one shared RuntimeNameSchema (42 characters, the memory-name cap less "Memory") behind both --name and the wizard.
The wizard asks only what scaffolding a runtime needs — name and template. --description still sets one on the flag path.
Wizard takes a doneLabel so the success footer can name what enter does; add runtime navigates back instead of exiting, the way build and deploy do.
An add screen the user navigated to should not tear the TUI down over a rejected name: onError=retry reports the failure in the wizard and returns to the form with the answers intact. project create keeps exit.
notgitika
force-pushed
the
feat/wizard-shell
branch
from
September 14, 2026 15:20
ea56301 to
9efa8ce
Compare
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
jariy17
previously approved these changes
Sep 14, 2026
agentcore-devx-automation
Bot
added
the
claude-security-reviewing
Contributor
|
Claude Security Review: no high-confidence findings. (run) |
agentcore-devx-automation
Bot
removed
the
claude-security-reviewing
jariy17
approved these changes
Sep 14, 2026
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
4 participants
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.
First of a few PRs. This one lands the shared wizard shell and puts two commands (agentcore project create and agetncore project add runtime) on it; the remaining
project addwizards follow separately.What
src/components/wizard/:Wizard: derives the step list from its<Step>children, owns position, key handling and the form → running → success | error phases, and renders the standard Layout + Stepper frame. Position is keyed by step name, so a{condition && <Step/>}branch appearing or disappearing does not move the user. A submit that streamsProgressEvents is rendered with the samedriveProgress+TaskListas create, build and deploy.Step: one question, one field.fields:TextField,ChoiceField,Summary. Each field owns its step'suseInputand publishes its own key hints, so a screen writes questions and nothing else.Callers
project createmoved onto it, unchanged in behavior. Its own controller (phases, step components, success panel, hint plumbing) is deleted; the two-level model step stays as a compound field, since the shell has no notion of focus within a step.project add runtimeis new: name → template → description → review.toAddRuntimeInputis extracted from the handler so the flag path and the wizard validate the same way.The runtime wizard scaffolds from a template. The JSON infrastructure flags,
--model-provider/--api-key(a secret read from stdin orfile://) and the Bedrock Agent import stay on the flag path.Shared components
The fields render existing components rather than new ones:
FormTextInput,FormRadioGroup,KeyValueTable,ErrorPanel,Layout,ui/{stepper,divider,spinner,task-list}. Two fixes at the source:KeyValueTablegrows into the space it is given, so its percentage key cap has a parent width to resolve against when the table is laid out as a row item.FormTextInput/FormRadioGrouprender bare when given no name or help text — the step already asks the question.Testing
bun test3218 pass / 0 fail;tsc --noEmit, oxlint and prettier clean. New:wizard.test.tsxfor the shell,runtime.screen.test.tsxfor the wizard (writes a real project and asserts the spec and scaffolded files),add.screen.test.tsxfor the add menu. Three help-viewport tests inproject.screen.test.tsxmoved toadd payment-manager, sinceadd runtimenow opens a wizard instead of its help.Next
HarnessWizard/EndpointWizardonto the shell (servesharness create/update,endpoint create/update, and givesproject add harnessa screen).