Ebbe BaßandClaude Sonnet 5.5
ae4b30ec22
Count and audit a wrong current password found at change time, log the unclear outcome with its details and end the other sessions as far as possible, note a lockout threshold of 1 in the audit, skip the local hints when the fine-grained policy cannot be read, round the minimum age up, reset the form after back navigation, give error messages unique ids, and fix the German rejection text. Adds tests for the block length, the masked password fields and the new cases; the review findings are recorded in the plan and the deferred work. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FocPJ75XPma7TcuYEMy8zm
3.2 KiB
3.2 KiB
- source_plan:
_bmad-output/initiative-ping-mee-portal/epic-platform-and-sign-in/story-audit-throttle-and-session-foundations-plan.mdsummary: Concurrent writers on the throttle counters are not tested. evidence: The test database is SQLite in memory with a shared cache, which answers "table is locked" to a second writer at once, so a threaded test fails for reasons that say nothing about production (a WAL file database with a 20 s busy timeout and IMMEDIATE transactions). A real test needs a file database; tried and dropped in the refactor sweep. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Code review 2026-10-19: the throttle check and the failure count are separate steps, so parallel requests can all pass the check before any failure is counted and more wrong binds reach AD than the limit allows. evidence: Pre-existing design of the throttle foundation (AD-7);portal/accounts/throttle.pycheckandhit. Needs a reserve-then-confirm step to fix, for every endpoint at once. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Code review 2026-10-19 (unverified, would be medium): a timeout after AD applied the password change is reported as "nothing changed", so sessions are not ended and no mail is queued. evidence: Settle by checking whetherLdapDirectory._service/modify_sturns a lost response intoDirectoryUnavailable, and whetherverify_passwordwith the new password after a timeout would tell. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Code review 2026-10-19: no signed-in page sendsCache-Control: no-store, so the back button can show a page after sign-out. evidence: Pre-existing; only the health check usesnever_cache. Decide once for every authenticated page (middleware), not per page. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Code review 2026-10-19: AD-2 in the architecture spine names amsDS-Lockout*wildcard, but the adapter reads onlymsDS-LockoutThresholdandmsDS-LockoutObservationWindow. evidence: Fix edits the architecture document; reword AD-2 to name the two attributes. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Known limitation, accepted in the code review of 2026-10-19: when AD never forgets failed attempts (a lockout observation window of 0), the portal's pause is still 15 minutes, so its limit does not keep up with AD's count. evidence: Unusual AD setup; not the owner's. Revisit if a deployment sets the observation window to never. - source_plan:
_bmad-output/initiative-ping-mee-portal/epic-account-self-service/story-password-change-plan.mdsummary: Known limitation, accepted in the code review of 2026-10-19: failed binds made elsewhere (authentik, other devices) count toward AD's lockout but not toward the portal's limit. evidence: Already stated in the plan's design notes. Fixing it needs a new directory read of the bad-password count (new adapter surface).