Skip to content

fix(monitor): guard the legacy goal's clear like the monitor's - #9845

Merged
iamwhatever merged 1 commit into
mainfrom
fix/clear-goal-confirm-and-danger
Sep 10, 2026
Merged

iamwhatever merged 1 commit into
mainfrom
fix/clear-goal-confirm-and-danger

Conversation

@bolichen97

@bolichen97 bolichen97 commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

PR #9788 merged with its UX review still at CONCERNS, and the finding is a real hazard rather than a style nit.

That PR gave a stopped automation an exit on both surfaces. Only one of them is guarded. On the monitor surface the erase is red, carries a trash icon, and asks first. In the legacy goal view the identical erase is one plain grey click that fires the DELETE immediately.

Blind reader, verbatim: "this delete button is plain grey with no warning colour, which makes it easier to hit by mistake" and "I cannot tell whether 'Clear stopped goal' deletes instantly with no 'are you sure?'".

Second finding, same review: neither confirmation row renders a question, so "Yes, clear it" answers one that was never asked. The reader: "If I hadn't seen shot-01 first, I wouldn't know what 'clear it' refers to."

Why it matters

The thing being erased is the goal text a user wrote, and there is no undo and no rebuild path. The control sits one slot left of the purple primary CTA, in the state a user is already frustrated in. A mis-click costs the whole goal.

What changed (motivation -> approach -> change)

Two surfaces, one operation, so they get one guard. The legacy control now:

  • is danger-coloured unconditionally, not on :hover. A touch viewport never produces hover, so a hover-only colour renders an irreversible erase identically to the buttons beside it (the same class of defect the Btn danger docstring records for Btn danger variant is hover-only, so destructive buttons look ordinary on touch #3937).
  • asks first, reusing the monitor surface's own two-step rather than inventing a second pattern.
  • has its confirmation replace the primary CTA instead of sitting beside it, so the row keeps the two-control cap (website/AUTOSDE.yaml:230), and the choice being confirmed is the whole row.
  • clears a pending confirmation on close, so reopening later cannot put a primed erase under the next press.

Both confirm labels now restate the action and its object: Clear goal for good and Clear monitor for good, across 12 locales plus the pseudolocale.

Tests

website/src/test/AutoNudgePopover.test.tsx:

  • the clear carries text-danger unconditionally.
  • one press erases nothing and offers the confirmation; Cancel restores the original control and still erases nothing; only the second press sends DELETE ?intent=clear.
  • the confirmation row holds exactly two controls, in order.

website/src/test/SessionAutomationPopover.test.tsx: updated for the relabelled confirm.

The capture harness gained a fifth frame and two assertions: that the first press sends nothing, and that the primary CTA is gone while confirming. It exits non-zero if either regresses.

Local: 90 tests across both popover suites, tsc -b, and npm run i18n:check all green.

Manual verification

N/A -- the capture harness drives the real built SPA and presses the control, so the two-step is exercised rather than described.

Screenshots / video

Stopped legacy goal -- the clear is red, and it asks:

the clear is danger-coloured

the confirmation replaces the primary CTA

The monitor surface's relabelled confirm

Clear monitor for good

Related Issues

Follows PR #9788, which merged before this review round landed.

Pattern harvest

Rule candidate: review-prompt
Pattern: two surfaces reach the same irreversible operation and only one is guarded. When a PR adds a destructive action to a second surface, the guard (confirm, danger styling, undo copy) is part of the operation, not part of the surface -- ask which surfaces can reach it and whether every one of them guards it the same way.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (N/A -- no behaviour a doc describes changed)
  • No secrets, credentials, or internal references in the diff

Round 1 review fixes (da0d40d13)

Lane Finding Fix
GPT 5.6 (blocking) a primed confirmation survived a websocket re-render, so another tab editing + restarting the same loop id (then a max_cycles=1 stop) left the press aimed at a goal the confirmation never described, with no 409 to catch it the confirmation resets whenever the loop's id, active state or message changes, not only on the open rising edge
First Principles (blocking) auto-nudge-clear and auto-nudge-confirm-clear had zero consumers both deleted; every selector uses role + name
First Principles (subtraction) the confirm row hand-rolled Btn's danger class string, which website/AGENTS.md forbids the three secondary controls are now <Btn> / <Btn danger>

The primary CTA beside them stays a raw button: it is pre-existing, no defect sits behind it, and swapping it to SendBtn would visibly change a control this PR has no reason to touch.

Negative control for the GPT fix: with the reset effect removed, the new test fails; restored, it passes. Full suite re-run after the change: 1979 files, 31305 tests, zero failures.

Round 2 review fixes (da0d40d13)

Lane Finding Fix
First Principles (concerns) the round-1 record-swap reset fixed only the legacy surface; SessionAutomationPopover holds two primed confirms (confirmStop, confirmClear) reset only on the open rising edge, and its press reads monitor.id at press time the same reset now guards both of them there, keyed on the monitor's id, active state, terminal outcome and target
UX (concerns) during confirmation the helper line still named the two buttons that had just left the row, on BOTH surfaces the line becomes the question for that state: "Remove this goal for good?" / "Remove this monitor for good?"
UX (concerns) the back-out read "Keep it" on one confirm row and "Cancel" on the other, for the same act one term on both: the goal row reuses the monitor row's own cancel string, and keep_goal is gone

The Korean strings for the two new questions use 하십시오체, which src/i18n/style/koStyle.test.ts pins for a destructive confirmation; the first draft used ~할까요? and that gate caught it.

Full suite after the change: 1979 files, 31306 tests, zero failures. The capture harness gained four assertions covering the question swap and the shared back-out word.

The one remaining evidence gap the UX lane named is a recording of the press-to-confirm swap. The transition is asserted instead, in two places that fail if it regresses: the vitest cases above, and the capture harness, which performs the press and checks that the first one sends nothing.

@bolichen97
bolichen97 requested review from a team and Zedmor September 10, 2026 03:11
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Sep 10, 2026
@github-actions

github-actions Bot commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Fable 5) — ✅ PASS

Design-level review of da0d40d136e2a055bcf760d99686a54783c759ba — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Real hazard (unguarded irreversible erase on one of two surfaces), closed by reusing the sibling surface's existing guard rather than inventing a new pattern.

Suggestions

  • The two-step confirm plus reset-on-record-swap now lives hand-synced in both popovers, and this PR's own review rounds already showed the drift (round 1 fixed one surface, round 2 copied it to the other) — a follow-up extracting a shared confirm hook/row would make the third divergence impossible.

[DESIGN-REVIEWED] da0d40d

@github-actions

github-actions Bot commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

Opus 4.8 Review — ✅ no blocking findings

Reviewed da0d40d136e2a055bcf760d99686a54783c759ba — this comment is updated in place on each push.

Review details

No findings.

The sole candidate is a transient one-frame flicker: an already-stopped loop with a primed clear-confirm that another client restarts renders, for one post-paint frame, "Stop loop" without the Save button before the reset useEffect (deps [loop?.id, loop?.active, loop?.message]) fires and restores it. The reachability depends on a cross-client restart landing precisely during a primed confirm, and the outcome is a sub-frame cosmetic flicker that self-corrects — it does not rise to a reachable, concrete defect at the required bar. Dropped.

[OPUS-REVIEWED] da0d40d

Verdict parsed from the review's SHA-scoped output markers for commit da0d40d136e2a055bcf760d99686a54783c759ba.

False positive or not applicable? A repository writer can comment:
/ai-review override fable da0d40d136e2a055bcf760d99686a54783c759ba: <one-sentence reason>

@github-actions

github-actions Bot commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Fable 5) — 🟡 CONCERNS

UX-level review of da0d40d136e2a055bcf760d99686a54783c759ba — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

UX-Verdict: CONCERNS

The guard reads correctly cold — every confirm label and question was understood — but the press-to-confirm row swap ships with no recording, only assertions.

Watch

  • The blind reader could not tell whether "monitor" (shot-01) and "goal" (shots 02–03) are one feature or two — "I cannot tell whether they are one thing shown two ways… that's the thing I'd most want explained" — and the now-identical "Remove this ___ for good?" / "Clear ___ for good" templates deepen the resemblance. Every user with both surfaces hits it, friction not failure, every encounter; a distinguishing clause in one of the two questions (or the popover titles) is the smallest fix, likely follow-up.

Evidence gaps

  • Lens-13 recording: pressing "Clear stopped goal" hard-swaps the row (confirmClear ? <Cancel/Clear goal for good> : <Btn danger>, Start loop unmounts) — reason stated in the PR, so not a block, but the harness's PNG frames and assertions cannot show the swap; a committed .webm/.gif of press → confirm → Cancel would close it (the author names this same gap in Round 2 notes).

Suggestions

  • stopped_help ("Clear stopped goal removes it for good.") still reads as an instant delete — the reader was "sure" one press erases and hesitated accordingly; add "Asks before removing." (and mirror in terminal_exits) so the guarded press isn't feared as unguarded.

[UX-REVIEWED] da0d40d

@github-actions

github-actions Bot commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Fable 5) — ✅ PASS

Premise-level review of da0d40d136e2a055bcf760d99686a54783c759ba — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

All claims verified: monitor.target exists, the two-per-row cap is a blocking AUTOSDE.yaml rule, Btn usage is a website/AGENTS.md invariant, the new i18n keys follow the 15-strong per-namespace cancel convention, and the hand-rolled two-step is meaningfully different from useArmedDelete (a same-button decay machine for row lists) while matching the monitor surface this PR is explicitly guarding against. The added tests fail on base, giving the defect checkable provenance. Final review:

First-Principles-Verdict: PASS

The 12 non-English confirm strings are new in this PR; only Korean style is machine-gated (koStyle.test.ts) — a human should spot-check the rest.

What this change ships

Inventory (10 items) — 10 justified

Intent: make the legacy goal view's irreversible erase ask first and look dangerous, as the monitor surface's identical erase already does — a FIX (UX review of #9788; the new tests fail on base).

  1. Legacy "Clear stopped goal" is danger-coloured always, not only on hover — justified
  2. The legacy clear now asks before deleting; first press sends nothing — justified
  3. The primary CTA leaves the row while confirming — justified
  4. A primed confirmation drops on close and when the loop record changes underneath — justified
  5. Monitor surface's two primed confirms also drop on record change — justified
  6. Monitor confirm relabelled "Yes, clear it" → "Clear monitor for good", 12 locales — justified
  7. Help/exits line becomes the question while confirming, both surfaces — justified
  8. Back-out reads "Cancel" on both confirm rows; keep_goal deleted — justified
  9. Three secondary controls become <Btn>/<Btn danger> — justified
  10. Capture harness fifth frame, new assertions, refreshed PNGs — justified

Notes a human can audit: item 3 is mandated by the blocking max-two-buttons-per-row rule (website/AUTOSDE.yaml:230); item 9 by website/AGENTS.md. Item 5 is the counted sibling of item 4's root cause (2 primed-confirm popovers grepped: AutoNudgePopover, SessionAutomationPopover) — the generality lens is satisfied, not violated. The two-step is not a second spelling of useArmedDelete (2 consumers: SchedulePage.tsx, HooksPage.tsx): that hook is a same-button arm-with-decay for row lists; this surface replaces the row with an explicit Cancel, the pattern its twin already ships. New catalog keys (clear_goal_for_good, cancel, clear_goal_question, clear_monitor_question) each have exactly one component consumer, the normal shape for i18n keys, and per-namespace cancel matches 15 existing instances.

[FIRST-PRINCIPLES-REVIEWED] da0d40d

@github-actions

github-actions Bot commented Sep 10, 2026 •

Copy link
Copy Markdown
Contributor

GPT 5.6 Review — ✅ no blocking findings

GPT 5.6 completed its review of da0d40d136e2a055bcf760d99686a54783c759ba and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] da0d40d

False positive or not applicable? A repository writer can comment:
/ai-review override gpt da0d40d136e2a055bcf760d99686a54783c759ba: <one-sentence reason>

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Sep 10, 2026
@bolichen97
bolichen97 force-pushed the fix/clear-goal-confirm-and-danger branch from f3caad5 to 9dfe9fb Compare September 10, 2026 04:15
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: action required A blocking check or review needs attention and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Sep 10, 2026
Follow-through on the UX review of #9788, which merged with the finding open.

Both surfaces now erase a stopped automation for good, but only the monitor's
was guarded: the legacy goal view's "Clear stopped goal" fired the DELETE on one
plain grey click. A blind reader named it: "this delete button is plain grey with
no warning colour, which makes it easier to hit by mistake" and "I cannot tell
whether it deletes instantly with no are-you-sure".

The legacy control is now danger-coloured unconditionally (not on :hover, which
a touch viewport never produces) and asks first, reusing the monitor surface's
two-step. The confirmation replaces the primary CTA rather than sitting beside
it, so the row keeps its two-control cap.

Both confirm labels now restate the action and its object, because neither
confirm row renders a question: "Yes, clear it" answered one that was never
asked. They read "Clear goal for good" and "Clear monitor for good".
@bolichen97
bolichen97 force-pushed the fix/clear-goal-confirm-and-danger branch from 9dfe9fb to da0d40d Compare September 10, 2026 05:13
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running labels Sep 10, 2026
@iamwhatever
iamwhatever merged commit a8512c7 into main Sep 10, 2026
67 of 73 checks passed
@iamwhatever
iamwhatever deleted the fix/clear-goal-confirm-and-danger branch September 10, 2026 06:15
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Sep 10, 2026
iamwhatever pushed a commit that referenced this pull request Sep 12, 2026
Reverts #5185 and its three follow-ups (#9788, #9845, #10190), which replaced
the composer's generic monitor/auto-nudge popover with a pull-request-only
"Bounded pull request monitor" form.

The composer trigger opens AutoNudgePopover again, so a session can arm a goal
loop for any target from the dashboard instead of being asked for a GitHub pull
request URL. The frontend automation-record layer #5185 introduced goes with it:
SessionAutomationPopover, MonitorRadar, the monitoring/automation normalizer and
its contract.json, the per-slot REST automation snapshots in ChatPage /
ChatSidebar / chatSlice / useWebSocket, and the sessionAutomationPopover catalog
block in all 13 i18n catalogs. The follow-ups' clear-a-stopped-record path goes
with it too, since both surfaces it guarded were part of that layer.

The monitor_watch MCP tool and the src/kiro_crew/monitoring structured-monitor
engine are untouched: agents can still arm, inspect and stop a bounded monitor.
Only the fields and routes #5185 itself added for the dashboard surface are
removed with it.

Conflict resolutions worth naming, all preserving post-#5185 unrelated work:

- The 13 i18n catalogs plus en.manual.json keep every key intervening commits
  added; only the sessionAutomation / bounded-monitor keys are removed.
- ChatSidebar keeps its pending_decision status row and ChatPage its extended
  dependency list; ChatSliceCoverage keeps the fetchSlots import a later commit
  started using.
- learn-cron-dashboard.md keeps the provider-evidence and legacy-cap paragraphs
  that landed after #5185, and drops the AutomationRecord SPA spec block.
- The babysit skill no longer points at the removed Clear stopped monitor
  control.
- The throwaway temp-screenshots evidence files from #9845 and #10190 are
  deleted rather than restored.
iamwhatever pushed a commit that referenced this pull request Sep 12, 2026
Reverts the dashboard layer of #5185 and its follow-ups (#9788, #9845,
with a pull-request-only "Bounded pull request monitor" form.

The composer trigger opens AutoNudgePopover again, so a session can arm a
goal loop for any target from the dashboard instead of being asked for a
GitHub pull request URL. The frontend automation-record layer #5185
introduced goes with it: SessionAutomationPopover, MonitorRadar, the
monitoring/automation normalizer and its contract.json, the per-slot REST
automation snapshots in ChatPage / ChatSidebar / chatSlice / useWebSocket,
and the sessionAutomationPopover catalog block in all 13 i18n catalogs.

The generic popover keeps the clear-stopped-record control #9788 and #9845
added: a stopped record is cleared with a confirm two-step that sends
DELETE ?intent=clear, so an agent-armed monitor that was stopped by a
session close or user stop still has an exit from the dashboard.

Nothing under src/kiro_crew/ or test/ changes. The monitor_watch MCP tool,
the src/kiro_crew/monitoring engine, the backend clear route and the
retention refusal are all as on main.

A committed capture harness, website/scripts/capture-goal-popover-generic.mjs,
opens the built SPA, clicks "Set a goal" and fails if the popover carries a
pull request URL field or a "legacy goal loop" call to action.
iamwhatever pushed a commit that referenced this pull request Sep 13, 2026
Reverts the dashboard layer of #5185 and its follow-ups (#9788, #9845,
with a pull-request-only "Bounded pull request monitor" form.

The composer trigger opens AutoNudgePopover again, so a session can arm a
goal loop for any target from the dashboard instead of being asked for a
GitHub pull request URL. The frontend automation-record layer #5185
introduced goes with it: SessionAutomationPopover, MonitorRadar, the
monitoring/automation normalizer and its contract.json, the per-slot REST
automation snapshots in ChatPage / ChatSidebar / chatSlice / useWebSocket,
and the sessionAutomationPopover catalog block in all 13 i18n catalogs.

The generic popover keeps the clear-stopped-record control #9788 and #9845
added: a stopped record is cleared with a confirm two-step that sends
DELETE ?intent=clear, so an agent-armed monitor that was stopped by a
session close or user stop still has an exit from the dashboard.

Nothing under src/kiro_crew/ or test/ changes. The monitor_watch MCP tool,
the src/kiro_crew/monitoring engine, the backend clear route and the
retention refusal are all as on main.

A committed capture harness, website/scripts/capture-goal-popover-generic.mjs,
opens the built SPA, clicks "Set a goal" and fails if the popover carries a
pull request URL field or a "legacy goal loop" call to action.
iamwhatever pushed a commit that referenced this pull request Sep 13, 2026
Reverts the dashboard layer of #5185 and its follow-ups (#9788, #9845,
#10190), which replaced the generic goal-loop popover with a
pull-request-only "Bounded pull request monitor" form.

The composer trigger opens AutoNudgePopover again, so a session can arm a
goal loop for any target from the dashboard instead of being asked for a
GitHub pull request URL. The frontend automation-record layer #5185
introduced goes with it: SessionAutomationPopover, MonitorRadar, the
monitoring/automation normalizer, the per-slot REST automation snapshots
in ChatPage / ChatSidebar / chatSlice / useWebSocket, the dead
api.monitorClear client method, and the sessionAutomationPopover catalog
block in all 13 i18n catalogs. website/src/monitoring/contract.json stays:
test/test_monitor_persistence.py reads it as the backend parity artifact.

Two fixes #5185 carried incidentally stay in the restored wiring. The
goal-loop seed and the autonudge_state frame fold channel slot keys
(slack:<ts>) to the dashboard filename form (slack_<ts>) through
website/src/utils/dashboardSlotKey.ts, matching _normalize_slot_key on the
backend, so a channel-bound loop lights its sidebar row. Every seed takes
its own generation, so two overlapping reconnect seeds cannot let the older
snapshot overwrite the newer one, and ChatPage's per-slot loop fetch is
dropped when a live autonudge_state frame for that slot lands first. The
popover keeps the viewport-fit classes (max-h / overflow-y-auto, stacked
number fields on narrow widths).

The generic popover keeps the clear-stopped-record control #9788 and #9845
added: a stopped record is cleared with a confirm two-step that sends
DELETE ?intent=clear, so an agent-armed monitor that was stopped by a
session close or user stop still has an exit from the dashboard.

The popover's fresh-state defaults change in two small ways: Max cycles
starts at 24 (0 still means no cap, and an existing loop's 0 is shown as
0), and the helper text says a Stop loop button appears while the loop runs
and that the stop-file name in the goal is filled in at start
(components.autoNudgePopover.stop_loop_hint, all 14 catalogs).

Nothing under src/kiro_crew/ or test/ changes. The monitor_watch MCP tool,
the src/kiro_crew/monitoring engine, the backend clear route and the
retention refusal are all as on main.

A committed capture harness, website/scripts/capture-goal-popover-generic.mjs,
opens the built SPA, clicks the composer trigger in three stubbed states (no
loop, active loop, stopped record plus its confirm row) and fails if the
popover carries a pull request URL field or a "legacy goal loop" call to
action. Its frames are committed under temp-screenshots/goal-popover-generic/.
iamwhatever pushed a commit that referenced this pull request Sep 13, 2026
Reverts the dashboard layer of #5185 and its follow-ups (#9788, #9845,
#10190), which replaced the generic goal-loop popover with a
pull-request-only "Bounded pull request monitor" form.

The composer trigger opens AutoNudgePopover again, so a session can arm a
goal loop for any target from the dashboard instead of being asked for a
GitHub pull request URL. The frontend automation-record layer #5185
introduced goes with it: SessionAutomationPopover, MonitorRadar, the
monitoring/automation normalizer, the per-slot REST automation snapshots
in ChatPage / ChatSidebar / chatSlice / useWebSocket, the dead
api.monitorClear client method and its stale test mock, and the
sessionAutomationPopover catalog block in all 13 i18n catalogs.
website/src/monitoring/contract.json stays: test/test_monitor_persistence.py
reads it as the backend parity artifact.

Three behaviours #5185 carried incidentally stay in the restored wiring.
The goal-loop seed, the autonudge_state frame and ChatPage's per-slot
listener fold channel slot keys (slack:<ts>) to the dashboard filename form
(slack_<ts>) through website/src/utils/dashboardSlotKey.ts, matching
_normalize_slot_key on the backend, so a channel-bound loop lights its
sidebar row and its live frames reach the open popover, and a monitor's
frame feeds the sidebar the same agent-turn accounting the REST seed does.
Every seed takes
its own generation, so two overlapping reconnect seeds cannot let the older
snapshot overwrite the newer one, and ChatPage's per-slot loop fetch is
dropped when a live autonudge_state frame for that slot lands first. When
that fetch fails, the popover says so and withholds Save / Start / Trigger
behind a Retry loading button instead of offering an empty editor whose
Start the existing loop would refuse. The popover keeps the viewport-fit
classes (max-h / overflow-y-auto, stacked number fields on narrow widths).

A structured monitor armed through the MCP monitor tools reaches the legacy
slot read as a reduced row with no message, banner or stop-file path. The
popover recognises both transports of that row (the REST read withholds
message; the live autonudge_state frame carries a monitor object, so
isReducedMonitorRow keys on either) and shows only what the backend will
honour:
a notice that an agent armed a bounded monitor, the schedule line, Stop
loop while it runs, and the clear-stopped-record two-step once it stops.
Save, Start and Trigger, which the backend refuses for that row with
structured_monitor_requires_monitor_api, are not rendered for it.
AutoNudgeLoop.message is optional to say so; the one other reader
(MembersPage's patrol row) falls back to an empty string.

The generic popover keeps the clear-stopped-record control #9788 and #9845
added: a stopped record is cleared with a confirm two-step that sends
DELETE ?intent=clear, so an agent-armed monitor that was stopped by a
session close or user stop still has an exit from the dashboard.

The popover's fresh-state defaults change in two small ways: Max cycles
starts at 24 (0 still means no cap, and an existing loop's 0 is shown as
0), and the helper text says a Stop loop button appears while the loop runs
and that {{STOP_FILE}} in the goal is replaced with a real file path at
start (components.autoNudgePopover.stop_loop_hint; the moustache is passed
as an interpolation value so i18next renders it literally). The three new
strings and the reworded hint are translated in all 11 human catalogs;
en-XA is regenerated.

Nothing under src/kiro_crew/ or test/ changes. The monitor_watch MCP tool,
the src/kiro_crew/monitoring engine, the backend clear route and the
retention refusal are all as on main.

A committed capture harness, website/scripts/capture-goal-popover-generic.mjs,
opens the built SPA, clicks the composer trigger in five stubbed states (no
loop, active loop, stopped record plus its confirm row, reduced monitor
row, failed slot snapshot), fails if the popover carries a pull request URL
field or a "legacy goal loop" call to action, and clips each frame under
2000px. Its frames are attached to the pull request, not committed. The two
harnesses that photographed the removed surface, capture-goal-clear-record.mjs
and capture-monitor-trigger-centering.mjs, go with it; the clear-record flow
is state 03 of the generic harness.
amadsalmon pushed a commit to amadsalmon/KiroCrew that referenced this pull request Sep 22, 2026
…otdev#9845)

Follow-through on the UX review of kirodotdev#9788, which merged with the finding open.

Both surfaces now erase a stopped automation for good, but only the monitor's
was guarded: the legacy goal view's "Clear stopped goal" fired the DELETE on one
plain grey click. A blind reader named it: "this delete button is plain grey with
no warning colour, which makes it easier to hit by mistake" and "I cannot tell
whether it deletes instantly with no are-you-sure".

The legacy control is now danger-coloured unconditionally (not on :hover, which
a touch viewport never produces) and asks first, reusing the monitor surface's
two-step. The confirmation replaces the primary CTA rather than sitting beside
it, so the row keeps its two-control cap.

Both confirm labels now restate the action and its object, because neither
confirm row renders a question: "Yes, clear it" answered one that was never
asked. They read "Clear goal for good" and "Clear monitor for good".
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants