Build and enable Slack invite/approval flow on the new website RC
メンテナーはふだん 1 日以内に返信
まだ誰も着手していません。
評価
- 難易度
- 5/5
- 見積もり時間
- 1週間以上
- 初心者へのやさしさ
- 15/100
- issue の種類
- 機能追加
- 明瞭さ
- おおむね明確
- 活発さ
- 活発
- 技術スタック
- python
調査の方向性
Start with _docs/specs/slack-invite-approval.md on the intake branch, then read the listed app-boundary, Studio/admin API, security, and selective-CI documents. The issue spans application workflows, staff capabilities, Slack join monitoring, deployment, and independent browser verification; the listed acceptance criteria and scenarios define completion. Check provider capabilities and deployment prerequisites before implementation, and report any unavailable capabilities explicitly.
索引モデルが issue の本文から書いたものです。
説明
Owner intake: be517170-c46d-41db-ac10-d0a4ebdcc924, points 1–6 and all notes t2545u (vouch), t2546u (RC enablement correction), t2547u (monitoring). The earlier no-rollout wording is superseded only for prod.datatalks.club; current live signup remains untouched.
Normative design: _docs/specs/slack-invite-approval.md on the intake branch. Relevant boundaries: _docs/architecture/app-boundaries.md, _docs/specs/06-studio-and-admin-api.md, _docs/specs/07-security-privacy-operations.md, _docs/ci/change-selective-ci.md. The new design supersedes older automatic shared-link Slack onboarding specifications for the RC.
Scope: database-owned site application/questionnaire, member invite/vouch accountability, staff approval and course auto-approval, invite graph and cascade site-access bans, one-use site redirect, join invalidation and read-only bypass monitoring; enabled RC deployment and controlled end-to-end verification. No Slack posts, real-person test invites/emails, current signup-email changes, join.datatalks.club changes, Slack settings changes, or automatic moderation by the monitor. Tests use only [email protected] recipients.
Staff management belongs to the existing capability registry and Studio/admin API parity contract. Declare five registered capabilities: application list query, application decision command, cascade website-access revocation preview query, cascade revocation confirmation command, and join report query. The separate preview must show the exact affected scope/count before the guarded confirmation. Each calls the same service through canonical Studio HTML and admin API adapters. Applicant/vouch/member self-service remains separate from staff management.
Acceptance criteria:
- Applicants enter who they are, where they work, and what they do and see private pending/approved/joined/rejected/banned status. Newsletter subscription remains separate.
- Active verified existing Slack members can invite or vouch through a shareable request link. The vouch page explicitly warns that they take responsibility with their own reputation, will also be banned if the applicant is a spammer, and should only vouch for people they really know. Confirmation requires authenticated POST and acknowledgement; banned/self/unverified sponsors are denied.
- Approval is on the site; authorized reviewers can approve/reject, course participants are auto-approved from actual course-domain evidence, and banned status overrides approval.
- Five staff capabilities (application list, application decision, cascade revocation preview, cascade revocation confirmation, join report) are registered with stable metadata and reachable through matched canonical
/studio/slack/HTML and/api/v1/admin/slack/adapters. Preview displays the affected scope/count; confirmation rejects changed state/scope/count. Positive/negative parity tests prove identical service behavior, explicit capability permissions, redacted audit, idempotency, concurrency, and applicable confirmation/session policy. The reviewedsite_adminmanifest includes the two Slack permissions and existingsync_studio_rolesapplies them without changing group members; no implicit superuser grant or owner waiver. Self-service/vouch routes do not acquire staff API authority. - Persist accountable inviter/voucher chains; a site ban revokes the member and all descendants atomically and blocks further sponsorship. Describe any actual Slack deactivation limitation honestly.
- Only approved unbanned applicants can redeem a bound opaque site redirect once atomically; GET/HEAD previews do not consume it. Joining Slack invalidates outstanding links. Raw Slack destination stays out of domain data, HTML/email/JSON/logs/reports; document unavoidable redirect-header visibility.
- Inspect actual existing dataops-managed token scopes and workspace plan without disclosing credentials. Programmatically ingest joins idempotently where current capabilities permit, retain observed inviter evidence when available, and report website-approved/outside/unresolved joins read-only. Preserve historical baseline separately. Document every plan/scope/provider restriction and whether ingestion/attribution are live; no automatic Slack action.
- Focused automated coverage proves cascade ban, single-use link, approval, vouch warning/acknowledgement and monitoring classification/no side effects; private responses are no-store.
- New management resource targets use stable UUIDs in both HTML/admin API adapters and OpenAPI, without integer-route fallbacks. GET/HEAD does not mutate application approval state; later course eligibility reconciles through an explicit CSRF-protected action or enrollment-owned mutation. New mutation forms use shared accessible fields/checks/actions, linked errors and error summary, with 44px control targets. Database-backed per-actor business limits return safe rate-limit responses, and repeated same-member vouch does not duplicate audit decisions.
- Independent tester inspects application, vouch, approval/report and denial/empty/error states at desktop and mobile sizes with screenshots; PM accepts after tester pass.
- PR checks are green on exact head before merge; deploy and verify enabled flow on prod.datatalks.club with a controlled test applicant. Preserve current join.datatalks.club, signup emails and workspace settings.
- Report at ~/tmp/agent-reports/2026-11-16-slack-invite-approval.md includes precise PR/commit/check/deploy/live evidence, actual provider capabilities/limitations, open questions for Alexey, and placeholder for Orin's separate banned-account findings.
Browser scenarios: new applicant submits complete questionnaire and gets pending state; known member opens shared vouch URL, reads warning, explicitly acknowledges and confirms; banned/self sponsor gets safe denial; course participant gets approved without sponsorship; authorized reviewer approves/rejects pending request; approved applicant redeems once and replay fails; join observation invalidates unused link; operator sees outside-flow and unknown-inviter rows without monitor-driven moderation. Verify responsive layout at desktop/mobile and no private data/credential exposure.
Management scenarios: authorized staff reaches application list/decisions/report/revocation from Studio; corresponding bearer-scoped admin API operations produce matching results and safe denials. A principal lacking the capability cannot read private questionnaire/report data or mutate approvals/bans. Stale/replayed commands follow declared concurrency/idempotency policy and do not duplicate business side effects. Cascade confirmation shows expected scope; audit never contains raw questionnaire text, email or Slack destination.
Dependencies: existing verified account identity and current course-domain participation; existing provider credentials and deployment authority. Unavailable API capabilities must be reported as explicit gaps rather than guessed. Orin's analysis is independent and does not block bounded implementation.
- 主要言語
- Python
- スター
- 0
- フォーク
- 0
- 平均マージ
- 37分
- マージ済み PR(30日)
- 36
環境構築
- Dockerfile または Docker Compose ファイルあり
- プルリクエストのテンプレートなし
- コントリビューションガイドなし
はじめの一歩
- issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
- 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
- リポジトリをフォークし、ブランチを切って変更します。
- issue 番号を参照したプルリクエストを送ります。
DataTalksClub/website のほかの issue
-
bug infra operations P1 testing
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
DataTalksClub/website#329 ·
メンテナーはふだん 1 日以内に返信
-
bug content data-migration documentation events operations P1
難易度 2/5 1〜3時間 初心者へのやさしさ 84/100
DataTalksClub/website#327 ·
メンテナーはふだん 1 日以内に返信
-
bug frontend P0 testing
難易度 2/5 1〜3時間 初心者へのやさしさ 90/100
DataTalksClub/website#300 · コメント 7 件 ·
メンテナーはふだん 1 日以内に返信
-
bug data-migration events P0 testing
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
DataTalksClub/website#295 · コメント 6 件 ·
メンテナーはふだん 1 日以内に返信
-
Show pending evaluation instead of premature project failure and zero scores (CMP drift)対応中かも このイシューにリンクされたプルリクエストがオープン中、またはマージ済みです。 オープンbug courses frontend P1
難易度 4/5 3〜5日 初心者へのやさしさ 35/100
DataTalksClub/website#525 · コメント 2 件 ·
メンテナーはふだん 1 日以内に返信
DataTalksClub/website の issue をすべて見る
似ている issue
-
needs-triage
難易度 1/5 1時間未満 初心者へのやさしさ 88/100
flashinfer-ai/flashinfer#9803 ·
メンテナーはふだん 1 日以内に返信
-
bug frontend good first issue help wanted
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
oscarbol09/thesisforge#61 ·
-
api bug performance
難易度 2/5 1〜3時間 初心者へのやさしさ 82/100
oscarbol09/audiobard#159 ·
-
area:web p2 product:business triage:active type:bug
難易度 1/5 1時間未満 初心者へのやさしさ 89/100
cuadrao-ai/argus#2328 ·
メンテナーはふだん 1 日以内に返信
-
bug frontend good first issue help wanted
難易度 2/5 1〜3時間 初心者へのやさしさ 88/100
nikhilb2/simple_invoicing#907 ·