Repository navigation
docs: record a kiro-wide uninstaller wipe as an accepted risk - #33699
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS Docs-only change recording a maintainer ruling. It takes nothing away that any code, release step or test reads. [DESIGN-REVIEWED] 877c2a2 Description read[DESCRIPTION-READ] 43724e276f18eea1fc8a2794fb7f03d17957bbbe93c8f07a2528057fa0a241c2 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 |
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: |
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: |
There was a problem hiding this comment.
Approving. Docs only: the UNINSTALLER-EXCLUDE-CREW "release gate" is rewritten as a documented accepted risk, following the maintainer's own decision on #355 (docs/system-specs/modules/config.md:374-392, docs/guides/install.md:1563-1574, docs/decisions/2026-11-09-a-kiro-wide-uninstaller-wipe-is-an-accepted-risk.md:1-30). Every factual claim checks out against source.
24a3b7a to
31f5e72
Compare
|
Rebased: main moved in docs/decisions/README.md, docs/guides/install.md and docs/system-specs/modules/config.md since this head went green (2caa69b -> fd5d9ef); CI re-running. GitHub reported a merge conflict in [operator: bolichen97] |
31f5e72 to
2e6be0a
Compare
|
Rebased: main moved in docs/decisions/README.md, docs/guides/install.md and docs/system-specs/modules/config.md since this head went green (fd5d9ef -> 0396be4); CI re-running. GitHub reported a merge conflict in [operator: bolichen97] |
buluoray
left a comment
There was a problem hiding this comment.
Approve on 2e6be0a. Since 31f5e72 the author only rebased the branch onto main 0396be4; the four-file docs diff is the same apart from the README index row's surrounding context. The accepted-risk decision is recorded by maintainer Bolin Chen and matches his #355 comment. Every source claim still holds at the new head. Decisive lines: docs/decisions/2026-11-09-a-kiro-wide-uninstaller-wipe-is-an-accepted-risk.md:3, docs/system-specs/modules/config.md:374, docs/guides/install.md:1588, src/kiro_crew/snapshot.py:254, src/kiro_crew/config/paths.py:781, src/kiro_crew/config/paths.py:1351
The data-home spec called UNINSTALLER-EXCLUDE-CREW a release gate that the product owner must sign off for every release, and the install guide called it release-blocking. Releases have shipped the data home under ~/.kiro/ since v0.1.1 without that sign-off, and nothing in this repository can constrain another product's uninstaller. Restate it as a documented accepted risk, list the protections Kiro Crew owns (its own uninstall paths keep the data home, the breadcrumb, KIROCREW_HOME, snapshot_dir outside ~/.kiro/), tell users in the install guide how to keep the data home and snapshots out of reach and that ~/.kiro/mates does not move with KIROCREW_HOME, and record the maintainer decision in docs/decisions. Closes #355
2e6be0a to
877c2a2
Compare
|
Rebased: main moved in docs/decisions/README.md, docs/guides/install.md and docs/system-specs/modules/config.md since this head went green (0396be4 -> ea742a7); CI re-running. The reds on the previous head were main's, not this PR's: the same Backend Tests node ids ( [operator: bolichen97] |
Problem / Motivation
Goal: The data-home spec and the install guide say truthfully that a Kiro-family uninstaller removing all of
~/.kiro/is a documented accepted risk, not a release gate, and tell users how to keep their data and snapshots out of its reach.The data-home spec calls UNINSTALLER-EXCLUDE-CREW a release gate: the product owner MUST confirm, for every release, that Kiro-family uninstallers skip
~/.kiro/crewor prompt first, and the install guide calls that sign-off release-blocking. Releases have shipped the data home under~/.kiro/since v0.1.1 without it, no release step reads it, and nothing in this repository can constrain another product's uninstaller.Why it matters
A spec that names a release gate nobody runs reads like a protection users have. The exposure itself (a
~/.kiro/-wide removal deleting the data home) is the same either way. What users can act on is the protections Kiro Crew owns, and today the install guide does not tell them to keep the data home or their snapshots outside~/.kiro/.Not a goal
~/.kiro/crewor prompts.What changed (motivation → approach → change)
A maintainer ruled on #355 that the gate becomes a documented accepted risk, so the two docs that called it release-blocking now say what is true and point users at the protections that exist.
docs/system-specs/modules/config.md, Data Home Location~/.kiro/wholesaledocs/guides/install.md, Windows NSIS paragraphKIROCREW_HOMEandsnapshot_diroutside~/.kiro/docs/decisions/The four protections are the repository's own uninstall paths (they keep the data home), the
~/.kirocrew.breadcrumbpointer,KIROCREW_HOME, andsnapshot_diroutside~/.kiro/. The name UNINSTALLER-EXCLUDE-CREW stays so a search still finds it. The install guide also says thatKIROCREW_HOMEstarts a new data home rather than moving one, that the crewmate agent files in~/.kiro/matesdo not move with it, and that a snapshot is not a full backup.Backwards compatibility
Compatible: docs only; no code, config, flag or test changes.
Removes nothing: no check or code path goes away;
git grep -n 'UNINSTALLER-EXCLUDE\|issues/355' origin/mainfinds only the two doc spots rewritten here, so no release step, script, skill or test read the old gate.Tests
No test: docs only.
scripts/docs_lint.pypasses (the new links and the#data-home-locationand#removing-user-dataanchors resolve), andscripts/check_decisions_history.pypasses (one new entry, no shipped entry touched).Manual verification
N/A — docs only. The facts the new text states were checked against the code:
snapshot.py_default_snapshot_dirfalls back toconfig_dir() / "snapshots";config/paths.pyconfig_dir()only creates aKIROCREW_HOMEfolder and copies nothing into it;v0.1.1is the first stable tag containing the data-home move.Related Issues
Closes #355
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)