Repository navigation
[release] v0.128.7 - #8106
Merged
Merged
[release] v0.128.7#8106
Conversation
added 30 commits
November 3, 2026 21:32
…rom prompts, variants, evaluators and traces
…yground selection
A late test set restore no longer connects after the page is left, chat rows wait for the revision's mode, a link into the open playground sets it up again, and trace spans follow the package's rules for span type, prompt inputs and evaluators.
Each set-up has its own token, so a restore that outlives it cannot touch the next set-up on the same revision; the address waits for evaluator drafts too; only a chat span is refused.
…y change of the address
…ables in /m
The snippet read the revision's input schema, which a prompt app often leaves
empty, so it sent "inputs": {} and a copied call failed with HTTP 400. Read the
input ports instead, as /w does, so each prompt variable gets add_a_value.
The list read the type from the app list flags, which carry no chat flag, so every prompt showed Completion. Read the latest revision's type, as /w does.
- An empty deploy note now stays empty, as in /w. /m sent undefined, and the package then wrote the raw "Deploy <app>.revision" key as the note. - The revision drawer title is Revision, and the delete toast says revision. - Modified by shows the member name. /m never binds the shared user atoms, so the author label could not resolve an id; the tab now passes the names from the workspace members.
…traces app filter The screen wrote the app scope atoms during its render. With the traces filters already mounted, that write updated TracesFilters while ObservabilityScreen rendered, and React logged an error. A pick in the toolbar now writes the scope in the change handler, so the render after the URL change has nothing to write.
The story still passed inputSchema, so the module threw on load. Also rename the builder's local map, so inputs means one thing there.
An app moved into a folder could not be moved back out. The workflow edit dropped a null folder_id (exclude_none), so there was no way to clear it. edit_workflow now keeps a folder_id the caller sent, including null. edit_application and edit_evaluator pass folder_id on only when their caller sent it, so a rename never moves an app or an evaluator. /m: the Move dialog gets an All prompts row that sends folder_id null.
…den-views-preview # Conflicts: # web/mobile/src/features/observability/ObservabilityScreen.tsx # web/mobile/src/features/variants/DeleteRevisionDialog.tsx # web/mobile/src/features/variants/DeploymentsTab.tsx
The release branch brought in three test files that fail the mobile lint on import/order. Sort the imports so the lint has no errors.
…ew-playground-links # Conflicts: # web/mobile/src/features/variants/VariantsTable.tsx
…den-views-preview
… what was not added, add annotation filters to live evaluation
The trace drawer clears its params after the close animation. By then the Playground button has moved to the playground, and the push there dropped the snapshot with the trace inputs. Clearing is now a no-op when the address has no trace or span param.
…es, count archived revisions in the next version
…s; shorten the strict read option
… loads its own annotation row
…live query cannot apply
… set fields The trace annotate sheet matched annotations against the login uid, but an annotation stores the profile id as its author. A second open showed an empty form and a save created a duplicate annotation. Read the id from userAtom. The shared annotation form (run annotate and queue sessions) sent every schema field, so untouched optional fields were saved as "" or null. Send only the fields with a value, and keep a saved null answer the user did not touch.
…n page Add runScenarioInvocation to @agenta/playground and a Run button in the Annotate tab and the annotate sheet of a human run, so a run made in /m can be annotated without opening /w.
…d it in /m The /m app shell registered the playground auth-header provider through executionItemController from the @agenta/playground root, which pulled the whole execution runner into the shell of every page. Import only the leaf header atom through a new @agenta/playground/execution-headers entry, and register it at AppProviders module scope so it is set before any run.
Contributor
|
Important Review skippedBot user detected. To trigger a single review, invoke the
 |
fix(api): live evaluations fail every evaluator step and re-evaluate their own evaluator traces
perf(frontend): load the playground runner only on the pages that need it in /m
fix(frontend): report a saved app rename as a success when the read-back fails
Contributor
Author
Website previewPreview URL: https://pr-8106-agenta-website-preview.mahmoud-637.workers.dev Built from |
Contributor
Author
Railway Preview Environment
Dependent tests finished. Post /preview for manual review. Environment absence verified; retained storage requires separate provider accounting. Post |
…, so its cleanup does not cancel the navigation
[fix] Edit your own /m trace annotation and save only the set fields
fix(frontend): queue scan stops on a failed trace read, Retry sends only the rest, live filter drops annotation fields
fix(frontend): /m playground fixes before preview (phone header, JSON text type, commit version)
[feat] Run a scenario's invocation from the /m evaluation run page
fix(frontend): fix the endpoints snippet, prompts type, variants labels and the traces filter warning in /m
fix(api): let an app move from a folder back to the top level
Keep both import sets in usePlaygroundSelection.ts. The compare test now allows the scroll option on the address write.
feat(frontend): Playground share links, last selection and links in from other /m pages
fix(frontend): show a plain message when the /w queue add fails
mmabrouk
approved these changes
Nov 5, 2026
mmabrouk
left a comment
mmabrouk
left a comment
Member
There was a problem hiding this comment.
🤖 The AI agent says: Approved on Mahmoud's instruction after the v0.128.7 production deploy went green in US (run 42008622294) and EU (run 42010227367), and the live checks passed in both regions.
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
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.
New version v0.128.7 in
Final QA
Stage QA ran on GKE staging and oss with api, workers and cron
92343f5e, web and web-mobileccdb0cd2(after the #8112 fix), services and runner5594f5ca.folder_id: nullreturns 200).Deploy steps
5594f5ca.