Repository navigation
feat(privacy): offer Turn on once to installs off only by an old written false - #37463
Conversation
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Nothing blocks: the owner decision covers asking these users once, the probe fails closed, "Keep off" records the acknowledgement, and the rendered evidence covers the new off form. Notes
[DESIGN-REVIEWED] ba482ab Description read[DESCRIPTION-READ] b27caa449f8c2b23339f70ac8523999c19e87b822de8c4d774e31da19ec9fd2d That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing blocks. The blind reader read both primary buttons correctly, and every new state is in a screenshot. Notes
[UX-REVIEWED] ba482ab Description read[DESCRIPTION-READ] 04ed404bf2ef0517a40f603340c6146781707da7a13cf6637eebc962a89daf2e That sha256 names everything this verdict read: the PR description, plus the 9 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
First Principles Review (Opus 5.5) — ✅ human override accepted@chenmingwei23 overrode this lane for |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
c44f65b to
ad2b84a
Compare
|
Product decision (owner) for head ad2b84a: Builds before the default-on change wrote a full default config on first start, including |
|
/ai-review override first-principles ad2b84a owner-decision: owner decided a one-time off-form Privacy prompt (and one CLI line) for installs off only by an old written false; default stays off: #37463 (comment) |
Human judgment recorded@chenmingwei23 recorded a owner-decision release (class
This decision applies only to this commit. A new push requires a new judgment. |
ad2b84a to
298f647
Compare
298f647 to
c748b8c
Compare
Code change approver recommendation: A person should decideWhy: the decider thinks the change is probably fine but estimates a 20% chance that this change introduces a defect, above the 12% limit for approving on its own. The PR adds a one-time Turn on / Keep off offer for public-build installs that are off only because an older build wrote telemetry.enabled:false. off_offer_due fails closed and excludes env pins, DNT, the heartbeat being off (so CLI and dashboard disablers are excluded), the overlay, policy, LTS, CI, non-default homes, the user's own collectors and an already-acknowledged notice. Each exclusion has a test. Turn on writes the switch, checks it took effect, then records the Kiro acknowledgement; Keep off records only the notice version. Kiro export stays gated on kiro_telemetry_ack_version, so no path sends data without consent. Failure paths read the state back and are tested. The theme payload and firstRun changes are consistent with existing state readers. The only red lane is the per-commit First Principles process gate, and the owner decision behind it is recorded. Approve, pending the CI jobs still running. An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit |
4e70bbb to
1c05e7e
Compare
…ten false Builds before local recording was on by default wrote telemetry.enabled: false into every new config, so most existing public-build installs are off without having chosen it and never see the version 3 notice that names Kiro telemetry. kiro_export.off_offer_due names that one case: a written false, the heartbeat on (including no KIROCREW_TELEMETRY_DISABLED), no env pin (DO_NOT_TRACK, KIROCREW_TELEMETRY), no policy pin, no config.local.json entry, no collector of the user's own, the destination listed (not LTS, not CI), the default data home (checked on its own, so dev homes that export are still not asked) and notice version 3 not yet acknowledged. For it the theme boot payload reports the stored notice version, and the collection status reports off_offer, so the dashboard shows the Privacy chapter once in an off form that names Kiro telemetry with Turn on performance metrics / Keep off. Its intro says turning them on also sends product usage metrics to Kiro telemetry, and that while off only the daily heartbeat and official app install receipts are sent; Turn on writes telemetry.enabled: true, confirms it took effect, then records the notice and Kiro acknowledgements. Any failure is followed by a read-back: recording still off says nothing changed; on says metrics are now on and the chapter shows the on form (Keep on / Turn off); unreadable says Turn on may now be on, or for Keep off that the choice could not be saved or confirmed (no button named that is not on screen). The off form lists the five heartbeat fields under the intro, before either button, and its body fades at the bottom while more is below. Keep off writes telemetry.enabled: false (already stored) and records only the notice version, so it is not asked again. A start with no dashboard prints one line pointing at `kirocrew telemetry enable`, once, on a terminal, and changes no setting. The theme payload and the telemetry startup state, which now read the policy profiles, run off the event loop. README, the export guide, metrics.md and the crew-log telemetry RFC describe the exception.
1c05e7e to
ba482ab
Compare
|
/ai-review override first-principles ba482ab owner-decision: owner decided a one-time off-form Privacy prompt (and one CLI line) for installs off only by an old written false; default stays off: #37463 (comment) |
Human judgment recorded@chenmingwei23 recorded a owner-decision release (class
This decision applies only to this commit. A new push requires a new judgment. |
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
I reviewed head ba482ab. off_offer_due fails closed. It skips DO_NOT_TRACK, the env pins, the heartbeat being off, a policy pin, a config.local.json entry, a collector of the user's own, LTS, CI, a non-default home, and a notice that is already acknowledged. Each of those has its own not-due test, and deleting any one of those checks makes at least one test fail. Turn on writes telemetry.enabled: true only after a click. Kiro export still waits for its own acknowledgement. The new config reads in the theme payload run off the event loop. PR Readiness is green on this head.
One thing done well: Turn on reads the collection status back before it records any acknowledgement, and every failure path reads the state back before it tells the person anything.
The design is signed off under the standing maintainer sign-off for changes by team members.
- Optional: Keep off only writes
privacy_ack_version, andoff_offer_due(kiro_export.py:245) and the CLI stamp check (notice.py:191) compare it against the livePRIVACY_NOTICE_VERSION. So the next notice bump asks everyone who chose Keep off again. You could compare against a fixed version 3, or store Keep off under its own key, and add a test that bumps the constant. - Optional: no test fails if you delete the
off_offer_dueguard in_maybe_print_offer(notice.py:193). A not-due case on the headless terminal path, reusing the table fromtest_it_is_not_due, would cover it. - Optional: the RFC and metrics.md hunks could say that someone who turned off only performance metrics in Settings is asked again, as the README already does. The CLI disable command and the chapter's Turn off are not asked again, because both also turn the heartbeat off.
…f-form offer (#37460, #37463, #37545) Backport to release/0.13.0 of three main commits, picked with git cherry-pick -x in main's merge order and squashed into one commit. They follow up #37183, which #37471 already put on 0.13: - #37460 (2a679d6): the built-in kiro-telemetry destination is also returned under a non-default KIROCREW_HOME. kiro_export.destinations() drops only the is_default_home() check; the LTS, CI and test-suite gates and every consent gate stay. The heartbeat keeps its own default-home rule. - #37463 (8de422d): installs that are off only because an older build wrote telemetry.enabled: false into their config are shown the Privacy chapter once, in an off form that names Kiro telemetry, with Turn on performance metrics / Keep off. kiro_export.off_offer_due decides that case; Keep off records only the notice version so it is not asked again. - #37545 (09b0ddc): test only. The TestOffOffer tests fake the default data home, since the offer is asked only there and the test home is a private KIROCREW_HOME. All three picks applied without conflicts; each backported patch is identical to its main commit (git range-diff shows no change). (cherry picked from commit 2a679d6) (cherry picked from commit 8de422d) (cherry picked from commit 09b0ddc) Co-authored-by: Raymond Chen <chenmingwei23@users.noreply.github.com>
…f-form offer (#37460, #37463, #37545) Backport to release/0.13.0 of three main commits, picked with git cherry-pick -x in main's merge order and squashed into one commit. They follow up #37183, which #37471 already put on 0.13: - #37460 (2a679d6): the built-in kiro-telemetry destination is also returned under a non-default KIROCREW_HOME. kiro_export.destinations() drops only the is_default_home() check; the LTS, CI and test-suite gates and every consent gate stay. The heartbeat keeps its own default-home rule. - #37463 (8de422d): installs that are off only because an older build wrote telemetry.enabled: false into their config are shown the Privacy chapter once, in an off form that names Kiro telemetry, with Turn on performance metrics / Keep off. kiro_export.off_offer_due decides that case; Keep off records only the notice version so it is not asked again. - #37545 (09b0ddc): test only. The TestOffOffer tests fake the default data home, since the offer is asked only there and the test home is a private KIROCREW_HOME. All three picks applied without conflicts; each backported patch is identical to its main commit (git range-diff shows no change). (cherry picked from commit 2a679d6) (cherry picked from commit 8de422d) (cherry picked from commit 09b0ddc) Co-authored-by: Raymond Chen <chenmingwei23@users.noreply.github.com>
Problem / Motivation
Goal: Public-build users who are off only because an older build wrote
telemetry.enabled: falsefor them are asked once, in the Privacy chapter, whether to turn performance metrics on; until they say yes, only the daily heartbeat (already on for them) is sent.Builds before local recording was on by default wrote a full default config on first boot, including
telemetry.enabled: false. #22397 reads any writtenfalseas the person's own opt-out, so most existing installs are off without ever having chosen it, and they never see the version 3 notice that names Kiro telemetry (#37183).Why it matters
Those people were never asked. Without this they stay off for good, and the product never hears from the bulk of its existing users, even ones who would say yes.
Not a goal
DO_NOT_TRACK, policy, heartbeat off, CI, LTS).What changed (motivation -> approach -> change)
Owner decision (2026-11-15): show these users one popup that names Kiro telemetry, with Turn on / Keep off (the button reads "Turn on performance metrics").
kiro_export.off_offer_due: a writtentelemetry.enabled: false, the heartbeat on, no env pin (DO_NOT_TRACK,KIROCREW_TELEMETRY), no policy pin, noconfig.local.jsonentry, no collector of the user's own, the destination listed (not LTS, not CI), and notice version 3 not yet acknowledged. Fails closed.off_offer;firstRun.tsxopens the chapter whenofferOn.PrivacyChapter.tsx: title "Performance metrics are off", an intro saying that turning them on also sends product usage metrics to Kiro telemetry and that while off only the anonymous daily heartbeat and official app install receipts are sent, the five heartbeat fields right under the intro, a line naming Kiro telemetry, a fade at the bottom of the body while more is below (off form only), and "Turn on performance metrics" / "Keep off". Strings in all 13 catalogs plus the pseudolocale.telemetry.enabled: true, re-reads the collection status to confirm it took effect, then records the notice and Kiro acknowledgements. Any failure is followed by a read-back. Recording still off: "nothing changed". On: "Performance metrics are now on, but your choice could not be fully saved", and the chapter shows the on form, so Turn off is the way back. Unreadable: Turn on says "... performance metrics may now be on"; Keep off says "Your choice could not be saved or confirmed". No acknowledgement is recorded on a failure.core.py: the three/api/theme/bootand/api/config/themeresponse sites load config and build the payload withasyncio.to_thread, sinceoff_offer_duereads config files and the policy profiles.telemetry.pyapi_telemetry_startupbuilds its state withasyncio.to_threadfor the same reason.telemetry.enabled: false(already the stored value, so it also undoes a Turn on that half-landed), then records onlyprivacy_ack_version, so it is not asked again.notice.OFFER_TEXTonce, on a terminal, pointing atkirocrew telemetry enable; it changes no setting.metrics.md, the crew-log telemetry RFC's "not asked again" rule.Backwards compatibility
No key, route or function is removed. The collection status gains
off_offer; clients that ignore it behave as before. Users off by any rule other than the writtenfalseabove see no change.Tests
test/metrics/test_kiro_export_destination.pyTestOffOffer: due and changes nothing; collection status reports it; not due for each of DNT, env pin,KIROCREW_TELEMETRY_DISABLED, heartbeat off, already acknowledged, unset, LTS, CI, a non-default home (checked on its own, so it holds after the dev-home export change in feat(metrics): send product metrics from non-default data homes too #37460), a collector of the user's own, aconfig.local.jsonentry, a policy pin; Turn on records and exports; Keep off sends nothing and is not asked again; the CLI line prints once on a terminal and writes no setting; no line into a log or on a dashboard start.website/src/components/PrivacyChapter.test.tsx: off form text, heartbeat fields and buttons; Turn on writes then acks; a failed write says nothing changed; a write that does not take effect says nothing changed; a lost response that landed, or an ack that fails after the write, says metrics are now on and shows Turn off; an unreadable read-back says may now be on; Keep off writes the switch off and records only the notice version. A Keep off whose write fails says nothing changed while still off, and a neutral could-not-be-saved line when unreadable.website/src/test/App.test.tsx: an existing user off only by the oldfalsegets the off form once.-n 2), frontend 230 + 139 style tests passed,tsc, pseudolocale check.Manual verification
Rendered with the built SPA on a loopback static server,
/apistubbed; screenshot below.Screenshots / video
Off form shown to a user who is off only because of an older build's written
false:The same form at 1280x900:
Turn on whose save fails (PATCH /api/config/kirocrew stubbed to 500): the read-back shows recording still off, so the form says nothing changed:
Turn on whose switch lands but whose acknowledgement fails (PUT /api/config/theme stubbed to 500): the read-back shows recording on, so the chapter shows the on form (title on, Keep on / Turn off) with "Performance metrics are now on, but your choice could not be fully saved. Keep them on, or turn them off.":
A failed write whose read-back is also unreadable (PATCH and the collection status stubbed to 500). The chapter falls back to its existing could-not-check form, which offers Turn off. Turn on:
Keep off (no button is named that is not on screen):
Related Issues
Follow-up to #37183. Related #21270.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.