Repository navigation
Conversation
Size Report
Startup median (7 runs, lower is better):
|
There was a problem hiding this comment.
1 issue found across 66 files
You’re at about 95% of the monthly reviewed-line limit. You may want to disable incremental reviews to conserve quota. Reviews will continue until that limit is exceeded. If you need help avoiding interruptions, please contact contact@cubic.dev.
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="website/docs/docs/browserstack.md">
<violation number="1" location="website/docs/docs/browserstack.md:11">
P1: `plugins add` fetches this package from npm, but `provider-browserstack` remains `private: true` and the release publisher rejects private tarballs, so this PR cannot publish what the guide tells users to install. Gate this instruction until a supported package source is available.</violation>
</file>
Reply with feedback, questions, or to request a fix.
View guided diff | Turn on auto-fix | Re-trigger cubic
| ## Install the plugin | ||
|
|
||
| ```bash | ||
| agent-device plugins add @agent-device/browserstack |
There was a problem hiding this comment.
P1: plugins add fetches this package from npm, but provider-browserstack remains private: true and the release publisher rejects private tarballs, so this PR cannot publish what the guide tells users to install. Gate this instruction until a supported package source is available.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. At website/docs/docs/browserstack.md, line 11:
<comment>`plugins add` fetches this package from npm, but `provider-browserstack` remains `private: true` and the release publisher rejects private tarballs, so this PR cannot publish what the guide tells users to install. Gate this instruction until a supported package source is available.</comment>
<file context>
@@ -3,7 +3,17 @@ title: BrowserStack
+## Install the plugin
+
+```bash
+agent-device plugins add @agent-device/browserstack
+```
+
</file context>
There was a problem hiding this comment.
This is retained as a publication prerequisite, rather than a temporary code workaround. The maintainer explicitly directed us to keep planned 0.22 install guidance and avoid temporary availability hardening. The package remains private; its placeholder, trusted publisher and usable publication must be verified separately. No claim is made that the package is already available, and the source repair at b41e2fe does not waive that release work.
There was a problem hiding this comment.
Recorded the requested explicit merge hold in the PR body: #4579 must not merge into 0.22 until supported installation delivers a usable BrowserStack package, the final packed live App Automate evidence is attached, and required CI passes. The package remains private until the maintainer placeholder and trusted publisher are verified; the existing install guidance is not evidence of availability. This publication thread remains open until those prerequisites are met. The iOS smoke failure is being handled separately, without a base-pass or infrastructure waiver.
Apex Sentinel · ✅ ApproveSince the last review: 0 new · 0 still open · 1 no longer raised (1 marked fixed) No open findings. Not in the reviewers' prompt (1 file)
Reviewed with this repository's own Sentinel configuration ( 💬 Talking to SentinelIn a finding's thread, reply with:
On this conversation, comment Earlier reviewsReview of
|
| Where | Finding |
|---|---|
packages/command-runtime/src/commands/interaction/runtime/interactions.ts:164 |
🆕 Temporary Settings-tap diagnostics ship in the shared tap hot path for all platforms |
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
1 finding below minor hidden (user_settings.min_severity).
Reviewed with this repository's own Sentinel configuration (83a8e530997f).
Review of 589c8d3f · 16.11.2026 11:31 UTC
Apex Sentinel · ✅ Approve
Since the last review: 0 new · 0 still open · 0 no longer raised
No open findings.
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
1 new finding away from what changed since 94489a2 not shown. Comment @apex-sentinel-staging review for a full review.
Reviewed with this repository's own Sentinel configuration (38e815742740).
Review of 94489a20 · 16.11.2026 11:11 UTC
Apex Sentinel · ✅ Approve
Since the last review: 0 new · 0 still open · 0 no longer raised
The changes are the same as on 09f643a, so this is that review, not a new one. Comment @apex-sentinel-staging review to review again.
No open findings.
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
Reviewed with this repository's own Sentinel configuration (ef19edc2a540).
Review of 09f643af · 16.11.2026 10:52 UTC
Apex Sentinel · ✅ Approve
Since the last review: 0 new · 0 still open · 0 no longer raised
No open findings.
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
Reviewed with this repository's own Sentinel configuration (ef19edc2a540).
Review of 85438c3d · 16.11.2026 10:19 UTC
Apex Sentinel · ✅ Approve
Since the last review: 0 new · 0 still open · 0 no longer raised
No open findings.
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
1 new finding away from what changed since b41e2fe not shown. Comment @apex-sentinel-staging review for a full review.
Reviewed with this repository's own Sentinel configuration (ef19edc2a540).
Review of b41e2fec · 16.11.2026 10:11 UTC
Apex Sentinel · ✅ Approve
Since the last review: 0 new · 0 still open · 1 no longer raised (1 marked fixed)
No open findings.
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
1 finding below minor hidden (user_settings.min_severity).
Reviewed with this repository's own Sentinel configuration (ef19edc2a540).
Review of 5bc427f1 · 16.11.2026 08:47 UTC
Apex Sentinel · ✅ Approve
1 finding(s): 1 minor.
🟡 Minor (1)
| Where | Finding |
|---|---|
packages/provider-browserstack/src/index.ts:3 |
Public options type requires both credential keys the code documents as optional |
Not in the reviewers' prompt (1 file)
pnpm-lock.yaml: generated or a lockfile, so its lines were left out
Reviewed with this repository's own Sentinel configuration (b96eb82b1b2a).
There was a problem hiding this comment.
Apex Sentinel: 1 new finding on 5bc427f. The verdict and the full list are in the summary comment.
|
I found two problems at 5bc427f to fix before merge. The provider scenario suite does not compile on this head. The diff removed the The PR body says the packed-plugin check and the live run were not done. This change is device-facing: lease allocation and session creation now go through the plugin registry instead of the bundled provider list. Nothing yet shows that the bundled dist loads from The open inline threads still apply: private package cannot be published (merging removes bundled support before the package can install), README plugin never passed, and loose smoke server, state directory wording, changeset daemon stop and credentials type, which are lower priority. The publishConfig.exports thread does not apply: Not blocking, take or leave: the factory in provider-browserstack/src/index.ts is a near copy of the TestMu one with the variable names changed, and The Fallow audit and the import-direction check also fail on this head. Fallow reports new exports, dependency and clone findings that overlap the new package and the copied factory. The import-direction check fails on the new provider-browserstack zone. I did not read their full logs, and I did not isolate the Coverage failure. The unused Before merge, please restore the import, clear the Fallow and import-direction failures, and add the packed check and live session evidence. The package also needs to be publishable before or with this merge. There are no conflicts. |
f8d3a6e to
e1a244d
Compare
5bc427f to
b41e2fe
Compare
There was a problem hiding this comment.
All reported issues were addressed across 13 files (changes from recent commits).
You’re at about 96% of the monthly reviewed-line limit. You may want to disable incremental reviews to conserve quota. Reviews will continue until that limit is exceeded. If you need help avoiding interruptions, please contact contact@cubic.dev.
Reply with feedback, questions, or to request a fix.
View guided diff | Turn on auto-fix | Re-trigger cubic
|
Thanks for the update. The review of 5bc427f (#4579 (comment)) is partly addressed: the README example, the smoke-test 404 path, and the Repo Guards fails in "Check import-direction DAG" (R11, package-boundaries.test.ts:837): This change is device-facing: BrowserStack lease allocation and session creation now go through the v2 plugin registry and the plugin's tsdown dist, and the bundled provider is removed. The last commits changed the built artifact again (shared build config, a second Not blocking, take or leave: Does this extraction need a new workspace package, a new external dependency (rolldown) and a new layering rank? Could the BrowserStack tsdown config keep the shape of provider-testmu's at the base? If the Fallow clone finding needs a shared owner, could Open threads: the thread on the package staying The large diff stat includes the rebase onto #4578; I reviewed only this PR's own commits. Before merge: declare rolldown 1.2.5 in the root dependency set (or drop build-config from this PR), get Repo Guards and Fallow green, make the package publishable, then post the packed check and the live packed-tarball session. |
09f643a to
94489a2
Compare
|
The actionable root-range finding is fixed in 589c8d3. Rolldown is used only by an erased type import, so its declaration and lockfile importer now use At that exact clean head, pinned pnpm11.17/Node26.8.1 frozen install and build succeeded. The requested command also exited 0: It packed the current checkout, installed core and plugin in isolated directories, checked the strict export map and absence of core/peer dependencies in the plugin installation, and exercised credential rejection plus connect against the local HTTP fixture. This is packed-contract evidence, not a live BrowserStack session. The checker removes its scratch tarballs after completion; no retained live-artifact provenance is claimed. The rank concern is resolved with its source rationale. The real App Automate session, video/log artifact evidence and placeholder/trusted-publisher prerequisites remain open. No temporary unavailable-package guidance was added. Hosted checks at older heads remain attributed to those heads; fresh CI is pending. |
|
The rolldown R11 fix from the earlier review (09f643a) is now in place. One problem from that review remains at 589c8d3, and the posted evidence is still short of what it needs. The BrowserStack help, docs tell users to run The open review thread on the same point still applies: #4579 (comment) For validation, please post one live BrowserStack App Automate session driven by the packed tarball at the final head. It should show Smoke Tests (ios) and the aggregate Smoke Tests fail on local-simulator Settings navigation. This diff does not touch that path, so they look unrelated, but I had no failure excerpt or base run to confirm. There are no conflicts. Before merge, the package must be publishable (or the hold recorded) and the live session must be posted. |
589c8d3 to
0c82e5f
Compare
There was a problem hiding this comment.
Apex Sentinel: 1 new finding on 0c82e5f. The verdict and the full list are in the summary comment.
…rowserstack plugin Move the BrowserStack provider modules and tests out of provider-webdriver into packages/provider-browserstack, packaged like the TestMu plugin (bundled ESM, type-only agent-device/plugins, ./agent-device export, version 2 factory). provider-webdriver keeps the shared engine and its ./plugin surface. Core no longer registers BrowserStack; connect browserstack without the plugin fails with the existing plugin-not-installed error naming plugins add @agent-device/browserstack. The package stays private until its npm placeholder and trusted publisher exist.
Document plugins add @agent-device/browserstack in the BrowserStack, device cloud and plugins guides, the README and the plugin package README, and add a changeset.
|
Two things from the 589c8d3 review are still open at 0c82e5f, and the code is not ready to merge yet. The other findings from that review no longer apply. The docs, README and help still tell users to run This PR moves provider registration for the production BrowserStack device route into the plugin. The packed checker against an HTTP fixture is the only proof so far, and no comment shows a live App Automate session from the packed tarball. Please install the packed The review thread on the private package still stands (#4579 (comment)), as does the debug probes thread (#4579 (comment)). Please remove the probes before merge, then confirm the iOS Settings smoke passes without them. Repo Guards & Compatibility fails only on gate daemon-wire-compat. This PR does not touch kernel, daemon-contracts, remote or contracts, and the drift comes from the 0.22 base, so it likely fails for any PR onto 0.22 until the base refresh in #4607 lands. Before merge, make the package publishable, post the live App Automate run, and remove the probes. |
0c82e5f to
7fce4ca
Compare
|
Thanks for the update. At 7fce4ca the probe commits are gone and the daemon-wire-compat failure is gone after the base refresh. All 21 checks pass and there are no conflicts. Two problems from the earlier review remain, and both need to be fixed before merge. The docs, README and help tell users to run The production BrowserStack provider registration moved from provider-webdriver into the plugin. The only proof so far is the packed checker against an HTTP fixture. Please install the packed The open inline thread on the unpublished package (#4579 (comment)) still stands. The thread about the probe commits (#4579 (comment)) no longer applies, because those commits were dropped and no probe code remains at this head, so please resolve it. Before merge, the package must be publishable (or the docs and help change held) and the live packed-tarball run must be posted. |
Summary
Extract BrowserStack into private
@agent-device/browserstackfor #3309 while retaining shared WebDriver behavior. Layer: 80 files, 2,142 gross lines. The eight extraction/build commits are preserved unchanged on the refreshed foundation #4607, now integrated into0.22at7fbdc887.All three temporary diagnostic commits are dropped. Both TypeScript Settings probes and native back-button logging are removed. The instrumented iOS observation passed and its probe artifacts are retained; the earlier failure’s cause remains unproven.
Merge hold: usable supported package installation and final packed live App Automate acceptance remain required. Keep
private: trueuntil the maintainer placeholder and trusted publisher are verified.Validation
Head
7fce4ca09149850716553812098ec662fddc2b5f, dependency7fbdc887.Node 24.13.0/pnpm 11.17.0 frozen install, build and full format passed.
exactCommit=true,checkoutChanged=false.The packed checker proves isolated exports, typed rejection and HTTP-fixture connect. It does not prove supported manager installation or live BrowserStack acceptance. Hosted CI at this head is green: 18 successful checks, two skipped and one neutral. Sentinel reports no open findings.
Open in [code]smith · Need help on this PR? Tag
@codesmith-botwith what you need.