[freeqaz]

Problem

Three browser test files added on 10-16/17 each hold one or two cases on a surface that an existing file already covers. Each costs its own Chromium launch, its own Console Browser 2 slot and its own suite entry.

Change

• console-create-agent-mode-defaults.test.mjs (test: pin conditions from the 10-17 night fixes #2703) moves into console-create-agent-provisioning-400.test.mjs, next to the other Create Agent execution-mode cases.
• console-create-agent-quota-access.test.mjs (test: pin conditions from the 10-17 morning fixes #2732) moves into console-create-agent-provisioning-400.test.mjs. The existing Pod quota loop is in console-agents-2.test.mjs, which open PRs fix: preserve native plugin permissions across Harness selection #2409 and fix: preserve native plugin permissions across Harness selection and admission #2258 edit, so it is not used here.
• console-agent-unsaved-secret-gate.test.mjs (test: pin conditions from the 10-17 morning fixes #2732) moves into console-agent-saves.test.mjs, after the other unsaved channel Secret gate cases. That file already defines the same UNSAVED_SECRETS_GATE and envBinding.
• checks-browser-2.json drops the three files.

Each moved test is copied verbatim and keeps its own fixture and setup (Sandbox Driver, default Presets, routed Installation answers, viewport). No test is renamed.

Left alone, because their home files are being edited elsewhere:

• console-agent-deploy-focus.test.mjs: its home, console-agent-draft-guards, is edited by feat: add per-agent sandbox network policies and denial review #2599 and feat: assign configured OpenClaw roles to Agent users #906.
• console-agents-list-return.test.mjs: being edited in another branch.
• console-refocus-focus.test.mjs: its home, console-retained-revalidation, is edited by feat: assign configured OpenClaw roles to Agent users #906 and another branch.

Verification

• Test names: the before set (3 removed files plus the 2 home files on main) and the after set (2 home files) are identical: 80 names each, including subtests, with no duplicates.
• Mutants, before and after: every mutant the moved tests killed is still killed in the new file, and the same moved tests fail.
  • Create Agent mode: C19, C20, C21, C22, C23, C31.
  • Pod quota line: Q1 (lost access), Q4 (403).
  • Deploy gate order: S3.
• node scripts/ci/run-tests.mjs audit: no issues. eslint and prettier are clean on the touched files.
• 20/20 at 2 CPUs on CPUs 10-13: console-create-agent-provisioning-400 passed 20/20 (36 tests each, 38-40 s per run) and console-agent-saves passed 20/20 (44 tests each, 44-49 s per run).
• Console Browser 2 lane time: the test window was 79 s on this PR's run (5b55cf1) against 82 s on main ac6a281, and 78-80 s on the four main runs before test: pin conditions from the 10-17 morning fixes #2732. That is the same within noise. The lane still runs 323 tests, now in 16 files instead of 19, and the summed file time fell from 496 s to 489 s. console-create-agent-provisioning-400 went from 29 s to 36 s and ends at 76 s, before console-agents-2 (79 s), which still sets the lane.

Risks

Test-only. In Console Browser 2, console-create-agent-provisioning-400 grew by 6 s in CI and still ends before console-agents-2. An internal review approved the change, with no changes requested. Its file-durations.json row (12.2 s) was already stale before this change, and it is not refreshed here.

[clawsweeper]

🦞👀
ClawSweeper picked this up.

Pull request received. I will update this pull request when review starts.

ClawSweeper review complete

ClawSweeper finished reviewing this revision. The review result is being finalized.

View the workflow run.

[freeqaz-openai]

Free's clanker here. I checked the actual base Console Browser 2 results (artifact 12098193551) and this PR's Console Browser 2 artifact (ID 12098143997). Both ran 323 tests with no failures or skips. All 80 cases in the five affected files match across the move; that comparison also rejects a changed quota assertion or a missing test replaced by a duplicate. The three moved test bodies are unchanged.

The concrete saving is 19 file runs down to 16, removing three normal Chromium launches. Tests keep separate browser contexts and app fixtures, but no longer get the three extra browser-process resets. I wouldn't claim a wall-clock speedup from this pair of runs. The comparison is for GitHub's tested merge 77396d46, not a claim that every later main revision was tested.

[clawsweeper]

Codex review: needs maintainer review before merge.

What this changes

This PR moves execution-mode, quota-access, and Secret-gate tests into two existing browser test files and removes their former suite entries.

Review scores

Measure	Result	What it means
Overall readiness	🐚 platinum hermit (4/6)	A focused, useful consolidation with preserved regression assertions, independent contexts, and passing relevant CI.
Proof confidence	🌊 off-meta tidepool	Not applicable: This maintainer-authored test consolidation changes no shipped behavior; relevant browser CI passes, and the supplied role-sharing screenshots concern a separate feature.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Product

Kind: Test only · Worth it: Yes · Fix scope: Complete
User problem: Contributors pay separate browser-process and CI scheduling overhead for small files covering existing Console surfaces.
Reason: The maintainer-authored change removes redundant setup while preserving distinct regression cases; elapsed-time improvement is not established or required to justify this small cleanup.

Merge readiness

✅ Ready for maintainer review

Keep this PR open: it preserves the regression cases while removing three separate browser-file runs, and I found no introduced defect.

Priority: P3
Reviewed head: 5b55cf1a5cad8b026fc795234922fbf8e0e1d75d

Before merge

None.

Findings

None.

Agent review details

How this fits together

The Console Browser 2 manifest selects browser test files; each file shares a Chromium process while its tests retain separate browser contexts and application fixtures.

flowchart LR
  A[Console Browser 2 manifest] --> B[CI test runner]
  B --> C[Two destination test files]
  C --> D[Shared Chromium per file]
  D --> E[Separate browser contexts]
  E --> F[Console and application fixtures]

Loading

Technical review

Best possible solution:

Land the focused consolidation with its existing regression coverage and browser-context isolation preserved.

Do we have a high-confidence way to reproduce the issue?

No product bug is claimed; source inspection confirms the separate-file overhead addressed by this PR.

Is this the best way to solve the issue?

The move reuses the existing file-level browser owner, preserves each test's setup and assertions, and removes the replaced registrations without adding another abstraction.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against 3d33ba8b51a0.

Provenance checked

• Execution-mode regression tests keeps the original intent (55659a6: Pin user-selected and credential-required execution modes while avoiding files under concurrent modification.)
• Quota-access and missing-binding gate regressions keeps the original intent (ac6a281: Pin denied deployment reads and missing-binding precedence after mutation testing exposed uncovered conditions.)
• File-level browser reuse keeps the original intent (3d9ad33: Reduce Chromium launch overhead while retaining independent contexts and fixtures; older follow-history traversal reached unavailable creation-adjacent blobs, without affecting inspection of this reuse contract.)

Testing

Proof path: in-process harness.

Security

None.

Evidence

What I checked:

• Introduced change preserves test bodies: The pinned introduction diff moves the two execution-mode tests, quota-denial test, and missing-binding gate test without changing their bodies; required imports and the default Preset load move with them. (5b55cf1a5cad)
• Browser isolation remains intact: fileBrowser shares only the Chromium process; newPage creates a new browser context and registers its cleanup with the application fixture, preserving cookies, routes, storage, and viewport isolation. (tests/browser/console-agents-browser-helpers.mjs:68, 5b55cf1a5cad)
• Current main still has the separate files: The six affected paths match the introduction base on fetched main, so the consolidation remains distinct work rather than an already-landed change. (3d33ba8b51a0)
• Relevant exact-head checks pass: Console Browser 2, Static Checks, and CI Required succeeded for the reviewed head in https://github.com/openclaw/openclaw-enterprise/actions/runs/39026913885; First Agent Smoke was still running. (5b55cf1a5cad)
• Recorded before-and-after comparison: The collaborator discussion at test(console): fold three single-case browser files into their homes #2736 (review) reports 323 passing tests without skips on both sides, with all 80 affected case names preserved; artifact metadata confirms the named base and PR runs. Raw job-log retrieval redirected to a blocked storage host, so those counts remain attributed to that discussion rather than independently re-counted. (5b55cf1a5cad)
• Original regression intent is retained: test: pin conditions from the 10-17 night fixes #2703 introduced execution-mode preservation cases, and test: pin conditions from the 10-17 morning fixes #2732 introduced the denied deployment-read and gate-order cases; both explain temporary separate-file placement around concurrent work. (ac6a281c61a9)

Review metrics

Metric	Value	Why it matters
Browser file runs	19 → 16	The manifest removes three files while retaining their cases in existing files, eliminating three normal Chromium launches.
Change footprint	Production +0/-0; tests +307/-360; manifest +0/-12	The patch removes duplicated file setup without changing product behavior.

Labels

Label changes:

• add P3: This is a narrow test-organization cleanup with no production behavior change.
• add rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• add status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.

Label justifications:

• P3: This is a narrow test-organization cleanup with no production behavior change.
• rating: 🐚 platinum hermit: Overall readiness is 🐚 platinum hermit; proof is 🌊 off-meta tidepool and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.

Rating scale

6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only. Repair and merge execution lanes are retired.

Reviewed October 17, 2026, 7:11 AM ET / 11:11 UTC.

[freeqaz-openai]

Free’s clanker here. Approved at 5b55cf1a. I compared the moved test bodies with current main: the assertions are preserved, the three old suite entries are gone, and the separately maintained focus test stays in place. The before-and-after browser results show the same 323 passing tests. The measurable improvement is three fewer Chromium launches, not a promised wall-clock speedup.

CI Required and Console Browser 2 passed. I read ClawSweeper’s complete review of this head; its “Before merge” and “Findings” sections are both empty.