fix(dashboard): restore isort-clean import order in the dashboard auth module - #2479
Conversation
Opus 4.8 Review — ✅ no blocking findingsReviewed Review detailsThe diff is a pure import-block reorder: the No findings. [OPUS-REVIEWED] 925d5fa Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 5.6 Review — ✅ no blocking findingsGPT 5.6 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 (Fable 5) — ✅ PASSAdvisory design-level review of The diff is a pure import reorder in one file — Design-Verdict: PASS Mechanical one-file isort fix that unbreaks a base-branch lint failure; zero behavior change, comments stay attached to their blocks. [DESIGN-REVIEWED] 925d5fa |
…h module (kirodotdev#2479) Co-authored-by: Joe Guo <zejiangg@amazon.com>
Problem
Backend Lint & Type Checkfails onorigin/mainitself, at theCheck imports (isort)step:
Because the job is a matrix, the 3.12 failure cancels 3.10, so the check reports two
red entries. It red-lights every open PR regardless of what that PR changes.
Why it matters
A base breakage on a blocking lint step makes every PR look broken and buries real
findings — the author's first move is to hunt for a fault in their own diff that is not
there.
Fix (symptoms → root cause → change)
Symptom: isort fails on a file no open PR touches.
Root cause:
d2e1109f9(#2388) added akiro_crew.dashboard.tailnetimport blockabove the
revocation_genblock, which is out of alphabetical order for that section.Change: run isort on the file, which moves the
tailnetblock belowrevocation_gen.Import statements only — no logic, no behaviour.
The two neighbouring blocks carry comments explaining that they exist to break a
token_auth↔refresh_tokensimport cycle. Each comment still sits directly above theblock it describes after the move, and the reordering cannot reintroduce a cycle:
tailnetdoes not import this module (its only in-repo imports aredashboard.urls,executors,platform.governance_profiles,platform_compatandsandbox), so itsposition among the absolute imports is inert at import time.
Tests
No new tests — this is an import-ordering fix on a lint-only failure, and isort itself is
the assertion. Verified on a clean worktree off
main(584bbb05f):isort --check-only src/kiro_crew test→ exit 0 (was exit 1)flake8 src/kiro_crew test→ exit 0Manual verification
N/A — the failing CI step is the verification, and it is deterministic. Import ordering
carries no runtime behaviour beyond module-load order, which is addressed above.