dashboard setup by aakrem · Pull Request #1 · Agenta-AI/agenta

aakrem · 2023-05-03T06:47:39Z

No description provided.

Format Python code with psf/black push

…ions ci/cd: added github actions for agenta-web

Get variant with env

…r `$ref=#/components/schemas/agenta_backend__models__api__evaluation_model__LMProvidersEnum-Input__1` see fix [#1](https://github.com/Agenta-AI/agenta_cloud/commit/3770adabaa1d0dd415db5ec18e4630e4f2dc1bcd)

- Add shouldIgnoreRowClick guard to AppTable row click handler - Add shouldIgnoreRowClick guard to VariantsTable row click handler - Fixes issue where clicking checkboxes or dropdown menus triggers row navigation - Affects app management page and all variant table contexts (overview, deployment, evaluation) Builds on shared utility from PR Agenta-AI#1

…ayground-entities-package-updates [Frontend] App playground local mode #1 - entities package update

JP shipped the TS Fern client with a default generators.yml ("0 seconds configuring Fern, maybe we need to re-generate"). Adds a config: block that addresses the bulk of the PoC's documented build/runtime issues at the source instead of via per-app workarounds. clients/scripts/generate.sh - omitFernHeaders: true — drops X-Fern-* headers the Agenta API CORS allowlist doesn't accept (PoC finding #5) - includeCredentialsOnCrossOriginRequests: true — withCredentials baked into every request so cookie-session auth works without a custom fetch wrapper (was finding in convenience layer) - fetchSupport: native, streamType: web, formDataSupport: Node18, fileResponseType: binary-response — prefer browser/web standards - retainOriginalCasing: true — keep wire snake_case (matches backend, OpenAPI spec, v2 entities Zod schemas; camelCase conversion would break ~all existing consumer code) - defaultTimeoutInSeconds: 30, maxRetries: 3 — explicit network defaults aligned with v2 stage-0 client - packageJson.browser: { fs/stream/buffer: false } + devDependencies: @types/node — bake the browser-stub and node types into the generated package.json (PoC findings #1 and #2) clients/typescript/package.json (bootstrap heredoc): mirror the same browser field + @types/node devDep so the static bootstrap matches. generate.sh adds fix_typescript_admin_duplicates: a Python post-gen patch that renames the second `createAccounts` (and its private __ partner) to `createAccountsAlt`. Two OpenAPI operations resolve to the same TS function name — backend should disambiguate via explicit operation_id. Admin endpoints aren't in v0 scope so the rename is safe. Serde stays OFF: noSerdeLayer: false + allowExtraFields exposed ~200 codegen errors in fern-typescript-sdk@3.63.7 (broken Record<string, T | null> handling, recursive type aliases for FullJson*/LabelJson*, plus the duplicate methods). Documented in generate.sh inline. The convenience layer's Zod boundary continues to handle Pydantic extra="allow" at the entity level instead. web/packages/agenta-sdk/src/index.ts: slim the fetch wrapper now that omitFernHeaders + withCredentials are upstream. The wrapper is only non-undefined when no apiKey was supplied — strips the empty Authorization header HeaderAuthProvider sets in the cookie-auth case. Verified: pnpm install clean; @agenta/sdk + @agenta/entities typecheck + lint pass; @agenta/client builds standalone (tsc with no errors); next dev --turbopack on web/ee returns HTTP 200 on /w (was 500 before the CORS fix) with no Module-not-found errors; next build on web/oss completes successfully.

After Arda's note that tracing works on main but not on our branch (suspected interaction between PR #4469 and our merge), did a full static-analysis pass on the FE tracing pipeline. Findings: * All tracing-related FE files are BYTE-IDENTICAL to main — `runnableSetup.ts` (references builder), the two tracing API endpoint files, `executionItems.ts` references-passing block, `store.ts` sourceRef setter. * All commits from PR #4491 (`fix/broken-tracing-and-workflow-events`, the fix JP shipped after PR #4469 broke tracing) are reachable from branch HEAD — `b1b5b899df`, `83495a340a`, `601c9de468`, `69379a9f04`, `301f74f65f`, `92be0a87e2`. * Our branch's actual diff vs main does NOT touch the trace pipeline, references building, sourceRef construction, or tracing endpoints. The diff is mustache token plugin, playground inputs body, visibility rule, template format picker, schema-aware seeding, view-type ordering, chat single- testcase gate, native-vs-stringified input transport. Recorded three plausible hypotheses in the inventory: 1. Premise re-check — "main works" may need re-verification with the same repro Mahmoud used. 2. Indirect effect from native-input transport — the only semantically meaningful FE change that touches the request body. Backend trace-save might choke on nested-object inputs even though trace_id was already returned. 3. Environmental — staging state, build cache, env var. Not code. Recommended next steps documented in inventory: re-verify main → capture wire-format diff → hand off to JP if backend-side.

…rk locally Confirmed via: * Local HEAD == origin HEAD (`396ecd5eda`). * `origin/main` (`d7c60c14e6`, frozen since 2026-06-01 10:21Z, i.e. 6 minutes BEFORE Mahmoud's tracing report) is an ancestor of our HEAD. * `origin/feat/add-mustache-rendering` (`a12751604a`, Arda's main- merge at 15:08Z) is an ancestor of our HEAD via merge `8dd4da1870`. * Static analysis earlier (commit `396ecd5eda`) confirmed all FE tracing files (`runnableSetup.ts`, `tracing/api/index.ts`, `trace/api/api.ts`, executionItems `payloadRefs` block) are byte-identical to main. * Arda re-tested locally on the synced branch and traces work. Net read: Mahmoud's #1 was a transient state at QA time, almost certainly the staging deploy catching FE post-#4469 / backend pre-#4491 (or the inverse) before JP's tracing fix had fully propagated. The code on our branch IS the post-fix state; staging behaves like local once redeployed. All nine inventory items now marked fixed/resolved.

After Arda's note that tracing works on main but not on our branch (suspected interaction between PR #4469 and our merge), did a full static-analysis pass on the FE tracing pipeline. Findings: * All tracing-related FE files are BYTE-IDENTICAL to main — `runnableSetup.ts` (references builder), the two tracing API endpoint files, `executionItems.ts` references-passing block, `store.ts` sourceRef setter. * All commits from PR #4491 (`fix/broken-tracing-and-workflow-events`, the fix JP shipped after PR #4469 broke tracing) are reachable from branch HEAD — `b1b5b899df`, `83495a340a`, `601c9de468`, `69379a9f04`, `301f74f65f`, `92be0a87e2`. * Our branch's actual diff vs main does NOT touch the trace pipeline, references building, sourceRef construction, or tracing endpoints. The diff is mustache token plugin, playground inputs body, visibility rule, template format picker, schema-aware seeding, view-type ordering, chat single- testcase gate, native-vs-stringified input transport. Recorded three plausible hypotheses in the inventory: 1. Premise re-check — "main works" may need re-verification with the same repro Mahmoud used. 2. Indirect effect from native-input transport — the only semantically meaningful FE change that touches the request body. Backend trace-save might choke on nested-object inputs even though trace_id was already returned. 3. Environmental — staging state, build cache, env var. Not code. Recommended next steps documented in inventory: re-verify main → capture wire-format diff → hand off to JP if backend-side.

…rk locally Confirmed via: * Local HEAD == origin HEAD (`396ecd5eda`). * `origin/main` (`d7c60c14e6`, frozen since 2026-06-01 10:21Z, i.e. 6 minutes BEFORE Mahmoud's tracing report) is an ancestor of our HEAD. * `origin/feat/add-mustache-rendering` (`a12751604a`, Arda's main- merge at 15:08Z) is an ancestor of our HEAD via merge `8dd4da1870`. * Static analysis earlier (commit `396ecd5eda`) confirmed all FE tracing files (`runnableSetup.ts`, `tracing/api/index.ts`, `trace/api/api.ts`, executionItems `payloadRefs` block) are byte-identical to main. * Arda re-tested locally on the synced branch and traces work. Net read: Mahmoud's #1 was a transient state at QA time, almost certainly the staging deploy catching FE post-#4469 / backend pre-#4491 (or the inverse) before JP's tracing fix had fully propagated. The code on our branch IS the post-fix state; staging behaves like local once redeployed. All nine inventory items now marked fixed/resolved.

…ub-block Two UX fixes in the agent config panel: - #1 Harness/model coherence: under the Model field, show a muted hint describing the harness's model convention (alias vs provider/id) and its providers, and warn when a configured provider isn't one the harness can reach. Non-destructive guidance (the Provider select is already gated); full model- list filtering is a follow-up needing provider-name normalisation against the model schema. - #2 Connection grouping: render Provider/mode/name as an inset bordered card with a muted CONNECTION sub-header, so the credential fields read as a sub-section rather than peers of Harness/Model.

…name refactor/agent-harness-rename renames the harness values pi→pi_core and agenta→pi_agenta (claude unchanged). Add pi_core/pi_agenta entries to the static capability map (both share Pi's exact slice: the eight vault providers, direct deployment, provider/id naming) while keeping pi/agenta as transitional aliases so the panel resolves against either backend during the rollout. The harness helpers and the #1 model-coherence hint key off these, so they now work for the renamed harnesses. Once feat/agent-model-picker lands the /inspect-fed map, the static fallback + aliases retire. Unit test covers the new names mirror the Pi set.

Reconcile the upstream agent work with our /invoke-absorbs-/messages direction. Ours wins on the contract: drop /messages, nest messages under data.inputs, use flags.stream. Keep upstream's batch/stream channel toggle (agentChannelModeAtom), pointed at /invoke and driving the Accept header (negotiation #1); x-ag-messages-format declares the message format (negotiation #2); history stays optional/default (#3). - WorkflowRequestData drops the messages and stream fields; reads move to data.inputs.messages and flags.stream. Alias kept. - is_agent inferred from the URI key at workflow creation; route gate removed. - Env vars normalized to our convention: AGENTA_AGENT_SKILLS_LOADED, DAYTONA_AUTOSTOP (code, helm, schema, docs, env examples). - Stale /messages comments in merged-in files updated to /invoke. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

ashrafchowdury

What looks good: Clean multi-stage Dockerfile with a non-root user, TypeScript strict mode enabled, Tailwind CSS properly wired, and the docker-compose service integration is well-structured overall.
Risky changes: (1) npm install --production in the Dockerfile deps stage skips devDependencies (TypeScript, ESLint, etc.) that npm run build requires — this will break the production Docker build. (2) The docker-compose volume mount ./dashboard:/app shadows the container's built-in node_modules, causing the dev server to fail unless node_modules is separately volume-excluded. (3) Placeholder API route with hardcoded data left in.
Missing tests: No test files or test script present anywhere in this PR. The package.json has no "test" script, and the new API route has no coverage.
Overall verdict: COMMENT — no critical security issues, but the Dockerfile production flag and volume-mount conflict are real bugs that will surface immediately. Please address before merging.

ashrafchowdury · 2026-07-08T11:29:59Z

+WORKDIR /app
+
+COPY package.json package-lock.json ./
+RUN  npm install --production


npm install --production omits devDependencies (TypeScript, ESLint, @types/*), but the builder stage copies these node_modules and then runs npm run build, which requires them. This will fail at build time.

Fix: use npm ci (or plain npm install) in the deps stage so all dependencies — including dev ones needed for compilation — are available to the builder:

RUN npm ci

ashrafchowdury · 2026-07-08T11:29:59Z

+    volumes:
+      - ./dashboard:/app
+    command: npm run dev
+    ports:


Mounting ./dashboard:/app will shadow the node_modules directory that was installed inside the image. The dev server will fail with Cannot find module 'next'.

Add an anonymous volume to preserve the container's node_modules:

volumes: - ./dashboard:/app - /app/node_modules

ashrafchowdury · 2026-07-08T11:29:59Z

+  req: NextApiRequest,
+  res: NextApiResponse<Data>
+) {
+  res.status(200).json({ name: 'John Doe' })


This is the default create-next-app placeholder route returning { name: 'John Doe' }. It should be removed before merging — it serves no purpose in a real dashboard and will confuse anyone hitting /api/hello in production.

The Agenta-AI#5190 follow-up: the elicitation form now has a real default field (shipped in Agenta-AI#5177), so the playbook layer retires its two workarounds — "put the recommended value in the field description" and "the FIRST enum option signals the default". - All 28 template playbooks migrated: proposed values become field defaults (one-click accept); researchable enums set "figure it out" AS the default and lean on the built-in Other… escape hatch; fields that already implied multiplicity use the multi-select array shape. Every body stays within its 1-2 KB budget. - write-template-playbooks skill: platform fact Agenta-AI#1 flipped (defaults exist — use them; type-matched, empty stripped, dates ignore them), skeleton + checklist + worked example updated. - playbook-spec: "The request_input reality (no defaults)" replaced with the shipped dialect (defaults, multi-select, oneOf choice cards); open-questions Agenta-AI#2 closed as shipped-and-adopted. - build-an-agent skill loop step 1 now tells the builder to propose defaults and use the richer shapes at the decisive prompt position.

dashboard setup

04cb14c

aakrem merged commit 9b137b6 into main May 3, 2023

aakrem deleted the AG-3-dashboard-read-data branch May 18, 2023 06:37

mmabrouk pushed a commit that referenced this pull request Aug 6, 2023

Merge pull request #1 from Pajko97/actions/black

7c90d73

Format Python code with psf/black push

aakrem pushed a commit that referenced this pull request Aug 31, 2023

Merge pull request #1 from lazyfuhrer/issue-471-agenta-web-github-act…

d2c17c3

…ions ci/cd: added github actions for agenta-web

mmabrouk pushed a commit that referenced this pull request Sep 19, 2023

Merge pull request #1 from yeokyeong-yanolja/get_variant_with_env

a4e8f73

Get variant with env

christseng89 added a commit to christseng89/agenta that referenced this pull request Oct 18, 2025

Setup Agenta Agenta-AI#1

9a9ad49

ardaerzin added a commit that referenced this pull request Dec 2, 2025

fixes latest eval runs tables feedback #1 & #2

757cb27

ardaerzin added a commit that referenced this pull request Feb 5, 2026

Merge pull request #3605 from Agenta-AI/frontend-feature/stateless-pl…

1399caa

…ayground-entities-package-updates [Frontend] App playground local mode #1 - entities package update

mmabrouk mentioned this pull request Feb 10, 2026

[fix] Resolve broken folders #3684

Closed

This was referenced Jun 19, 2026

docs(agent): agent-workflows ground truth + comment hygiene #4770

Closed

[docs] Add agent workflow interface inventory #4821

Merged

mmabrouk mentioned this pull request Jul 7, 2026

docs(design): agenta-skills sync review — gaps, priorities, implementation status #5108

Closed

ashrafchowdury reviewed Jul 8, 2026

View reviewed changes

mmabrouk mentioned this pull request Jul 9, 2026

feat(agents): Move agent-template intelligence from the card prompt into the build-an-agent skill #5188

Merged

ardaerzin mentioned this pull request Jul 10, 2026

feat(agents): template playbooks adopt real request_input defaults #5211

Merged

This was referenced Jul 12, 2026

feat(runner): deliver gateway tools to Claude on Daytona via an in-sandbox stdio MCP shim #5244

Closed

fix(web): honest sandbox picker + hide local-only permission options #5302

Merged

feat(runner): client tools on Claude in remote sandboxes #5383

Merged

mannietech15 mentioned this pull request Jul 22, 2026

(bug) Model picker shows a false "not available" badge, and Bedrock credential failures hang instead of failing fast #5118

Open

mmabrouk mentioned this pull request Jul 28, 2026

(bug) A session open in a second browser shows no sign it's running and gets permanently stuck on a stale transcript after refresh #5530

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

dashboard setup - #1

dashboard setup#1
aakrem merged 1 commit into
mainfrom
AG-3-dashboard-read-data

aakrem commented May 3, 2023

Uh oh!

ashrafchowdury left a comment

Uh oh!

ashrafchowdury Jul 8, 2026

Uh oh!

ashrafchowdury Jul 8, 2026

Uh oh!

ashrafchowdury Jul 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

aakrem commented May 3, 2023

Uh oh!

ashrafchowdury left a comment

Choose a reason for hiding this comment

Uh oh!

ashrafchowdury Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

ashrafchowdury Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

ashrafchowdury Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants