[freeqaz]

Problem

UX spec P5 (sections 3.5 and 3.6; plan rows S1b and D5). The Agent page shows its three facts (Selected version, Latest visible deployment, Live serving) as a two-column table. The sentences that say what the page cannot confirm are repeated under facts and cards: "Deployment records don't verify live serving or model access." under Live serving, "Selection does not confirm that this version is serving." in the current version's tab notice, and the listed-bindings caveat in the Share Agent panel.

Change

• Stat row (.stats inside a .card). There are three columns with hairlines between them. Each column has a 12 px muted label, an 18 px semibold value (--text-xl) and optional 12 px help. A recorded outcome is the version plus a .badge: Failed is danger, Completed is success, Queued and In progress are info, and Superseded and Stopped are neutral. Live serving is a badge too: "May be offline" is warning and "Not verified" is neutral. At 760 px and narrower the facts form one column, with each label and value on one line. The Runtime "Stop requested" fact, when present, takes a fourth column. The text stays "v7 · Failed": the separator is visually hidden, so screen readers and existing text matches keep it.
• One disclosure per page. A closed details.disclosure "What this page can tell you" (an underlined --text-md summary with a chevron) sits under the facts. It holds the three guarantee sentences verbatim. The bindings sentence shows only while the Share Agent panel is shown, through a CSS :has rule. The per-card copies are removed: the Live serving notes, the second sentence of the current version notice, and the bindings clause after "Current policy loaded."
• The deployment panel becomes its own card under the stat card, as in the mockup. The failed-deployment status block now has 32 px below it, which was the fable nit left from P4.
• The "Latest visible deployment" label is kept. The mockup says "Latest deployment", but "visible" says that only readable versions count. Renaming it would also reflow the agent-details.md table that five open PRs edit.
• Storybook: badge-and-card-workflow.md step 7 walks the stat row and the disclosure.

Verification

• New browser tests in console-agent-overview.test.mjs:
  • "The Agent's facts are a stat row with one disclosure" at 1440x900 and 390x844. It checks the structure and 18 px values, the badge tones, three columns or one column with label and value on one line, and that each guarantee appears exactly once on the page, inside the closed disclosure, and is visible only when it is opened.
  • "The page's disclosure names listed bindings only beside the sharing panel" (sharing read 403).
  • On main's sources all three fail (3/3). With only main's console.css, all three also fail (3/3).
  • Pinned 20/20 at 2 CPUs (3 tests per run).
• Touched and related browser files: 470/470 across 14 files. These are agent-detail, agent-detail-2, agent-logs, agent-overview, agent-saves, agent-state, agents-2, agents-list-return, retained-revalidation, console, status-block, button-hierarchy, page-header and deploy-again. After the final edits, overview, state and detail passed 101/101.
• After merging main (fix(console): call a deploying version deploying, not historical #2990, fix(console): name the default preset and fill use preset once a preset is chosen #2991, fix(console): open the configuration editor from edit current configuration #2992): 462/462 across the overview, state, detail, detail-2, deploy-again, saves, presets, retained-revalidation, agents-list-return, logs, console and agents-2 files.
• Visual review (internal):
  • At 761-1100 px, four facts overran their columns. Stat columns now take 16 px padding there, and a fourth fact (Stop requested) stacks the row. Probed at 761, 768, 860, 1024 and 1101: no badge or value overflow.
  • The stat and deployment cards are now 24 px apart.
  • The open chevron is nudged 2 px.
• prettier and eslint pass on the touched files.

Acceptance (spec 5, P5), measured with shots.mjs and a word probe on Storybook stories at 1440:

	main	this PR
.stats present	no	yes
Guarantee sentences visible outside the disclosure (draft / failed / succeeded)	1 / 2 / 3	0 / 0 / 0
Each guarantee in the DOM	0-1 copies, in cards	exactly 1, inside details
Words above the tabs, draft	294	280 (-5 %)
Words above the tabs, draft, without the Share Agent card (what P6 moves out)	145	144; vs 294 on main: -51 %
.button--primary on the failed page / accent / danger (shots)	2 / 3 / 6	2 / 3 / 7 (+1 is the red "Failed" badge, which the spec asks for)

The -40 % overview word target is shared with P6, which moves Share Agent (136 words) to an Access tab. P5 alone removes only the repeated sentences. The 330-word baseline in the spec is from dd92bd62a; main is at 294.

Open-PR conflict map (git merge-tree of this head and of origin/main against each open PR that touches detail.mjs, access.mjs, console.css, agent-details.md or scenarios.mjs):

• fix(console): name the default preset and fill use preset once a preset is chosen #2991, fix(console): open the configuration editor from edit current configuration #2992, feat: add Gateway managed mode to Agent deployments #2982, feat: share discoverable Driver configuration definitions #2980, fix(runtime): retry network, timeout and 5xx failures of the codex login and model check #2971, feat(console): show the authorized native Gateway people roster #2596, feat(console): stop and delete Agents from the Agents list #2524, Draft implementation: managed outbound HTTPS for dedicated Gateways #2518, Add OCE policy simulation and tester for network and plugin actions #2445, feat(auth): sign people in to occ with a device code (rfc-0019) #1619, feat(iam): grant the deployer exact read of the revision a deploy admits #1605 and feat: assign configured OpenClaw roles to Agent users #906 have the same conflict set as against main, so this PR adds nothing new for them.
• fix(console): call a deploying version deploying, not historical #2990 (versionNotice) landed while this PR was open. Merged main and resolved the one-line hunk: the current-version notice keeps fix(console): call a deploying version deploying, not historical #2990's Deploying branch and drops the guarantee sentence. fix(console): name the default preset and fill use preset once a preset is chosen #2991 and fix(console): open the configuration editor from edit current configuration #2992 also landed, with no conflict.

Media (secret gist): https://gist.github.com/freeqaz/78d0ec71d4a0eeeedc95aed0153e14ab

Failed deployment, 1440 (before / after / mockup):

Failed deployment, 390 (before / after / mockup):

Draft, 1440 (before / after / mockup):

Draft, 390 (before / after):

Disclosure open (1440 failed, 390 draft):

Recordings of the failed story at 1440 and 390, opening and closing the disclosure by click and then by Enter: disclosure-1440.webm · disclosure-390.webm

Risks

• This is a console-only layout and copy change; no API change. The guarantee sentences are unchanged, verbatim, and still on the page, folded.
• Relative time under "Latest visible deployment" (spec 3.5) is not added. The console has no relative-time helper, and the record's timestamp is already in Deployment details. It is left for a follow-up.
• The bindings sentence's visibility depends on CSS :has, which the console already uses for the one-primary rule.

[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.

[clawsweeper]

Codex review: needs maintainer review before merge.

What this changes

This PR presents Agent facts as responsive stats and consolidates repeated caveats into one expandable disclosure.

Example: Version v7 fails while v6 remains selected.

• Before: The overview presents the facts as table rows, with caveats repeated across cards.
• After: The overview shows v6, a red v7 Failed badge, and an amber May be offline badge above one disclosure.

Review scores

Measure	Result	What it means
Overall readiness	🦐 gold shrimp (3/6)	The focused patch has credible browser coverage and useful static visual evidence, but the repository's UI evidence requirement remains incomplete.
Proof confidence	🦐 gold shrimp (3/6)	Real behavior proof is not required for maintainer- or bot-authored pull requests.
Patch quality	🐚 platinum hermit (4/6)	No actionable review findings were identified.

Product

Kind: Feature · Worth it: Yes · Fix scope: Complete
User problem: Repeated caveats and table-style facts make the Agent overview harder to scan.
Reason: The maintainer-authored direction improves scanning while retaining the existing status meanings and warnings.

Merge readiness

✅ Ready for maintainer review

Keep open: this maintainer-authored PR makes a focused console improvement, and source review found no actionable correctness or security defect.

Priority: P3
Reviewed head: 0f9592f3d6bd276cd42d8b7b3a1a8862fdb32730

Before merge

None.

Findings

None.

Tests

• Missing end-to-end proof: The supplied evidence lacks the required final-revision video showing disclosure interaction at desktop and phone widths.

Agent review details

How this fits together

The Agent detail renderer consumes existing authorized Agent, deployment, and sharing-policy responses and produces console status cards without changing backend operations.

flowchart TB
  A[Authorized API responses] --> B[Agent detail renderer]
  B --> C[Responsive stat card]
  B --> D[Deployment evidence card]
  B --> E[Page disclosure]
  F[Sharing panel visibility] --> E

Loading

Technical review

Best possible solution:

Retain the focused presentation change while preserving the existing distinctions between selected, recorded, and verified runtime state.

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

This is a presentation enhancement; source comparison and the prepared before/after images establish the existing layout and intended change without executing target code.

Is this the best way to solve the issue?

The implementation uses the existing renderer, card tokens, badges, and native disclosure element while retaining permission-sensitive status and recovery logic.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against b225870963ba.

Provenance checked

• Agent overview and deployment evidence keeps the original intent (7c3ecbf: Separate current selection, readable deployment records, and live-serving uncertainty; later recovery work kept failures prominent.)
• Sharing-policy readback caveat keeps the original intent (48930e0: Current bindings cannot establish whether an earlier uncertain mutation settled.)
• Overview card layout changes intended behavior with a stated reason (4bf2532: Keep routine deployment evidence folded and failed-deployment recovery visible.)

Testing

Proof path: in-process harness.

Security

None.

Evidence

What I checked:

• Pinned introduced change: The introduced diff changes presentation, badge rendering, disclosure placement, workflow instructions, and browser coverage; it adds no API, permission, storage, or dependency changes. (apps/controller/src/console/agents/detail.mjs:1292, 0f9592f3d6bd)
• Current main retains the previous layout: Fetched main still renders the three facts using the previous summary structure, so this PR's central presentation change remains distinct. (apps/controller/src/console/agents/detail.mjs:1294, b225870963ba)
• Existing status contract: Clarify Agent deployment versions and add on-demand diagnostics #482 established separate selected-version, readable-deployment, and unverified-serving facts; Clarify Agent status and put failed-version recovery in the first screen #2542 preserved those distinctions while prioritizing recovery. This PR retains their status decisions. (apps/controller/src/console/agents/detail.mjs:2227, 0f9592f3d6bd)
• Sharing caveat provenance: feat: share an Agent with existing people #384 explains that policy readback reports current configuration rather than historical request settlement; this PR preserves that warning in the disclosure. (apps/controller/src/console/agents/access.mjs:261, 48930e064782)
• Visual evidence and applicable gate: Inspected all four prepared images, including the actual desktop after-image showing the new stats and closed disclosure. The PR supplies gist screenshots but no interaction video; AGENTS.md requires screenshots and a short video as native GitHub attachments. Additional gist retrieval failed with integration HTTP 403 and a sandbox domain block, which are review-environment limits. (AGENTS.md:454, 0f9592f3d6bd)
• Exact-head check status: The exact-head check response reports no failed checks and First Agent Smoke still running; this establishes no diff-caused CI failure. (0f9592f3d6bd)

Likely related people:

• Free Wortley: Raw commit 48930e0 adds apps/controller/src/console/agents/access.mjs:172 relative to its recorded parents. This identifies author metadata, not feature responsibility or a PR merger. (role: source-line author; confidence: high; commits: 48930e064782; files: apps/controller/src/console/agents/access.mjs)
• stevenlee-oai: Raw commit 7c3ecbf adds apps/controller/src/console/agents/detail.mjs:518 relative to its recorded parents. This identifies author metadata, not feature responsibility or a PR merger. (role: source-line author; confidence: high; commits: 7c3ecbfce4fa; files: apps/controller/src/console/agents/detail.mjs)

Review metrics

Metric	Value	Why it matters
Change distribution	Production +273/-76; browser tests +158; workflow documentation +8	Growth is concentrated in responsive presentation and observable browser coverage, with no new subsystem or dependency.

Labels

Label changes:

• add P3: This is console layout and readability polish with unchanged backend behavior.
• add rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦐 gold shrimp 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 console layout and readability polish with unchanged backend behavior.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🦐 gold shrimp and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.

Rank-up moves

Optional improvements that raise the rating; they are not merge blockers.

• Attach final-revision desktop and phone screenshots plus a short video opening and closing the disclosure.

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 19, 2026, 11:41 PM ET / October 20, 2026, 03:41 UTC.