plan-excalidrawSlidesPlan

Prompt file imported from angelator312/excalidraw-slides (.github/prompts/plan-excalidrawSlidesPlan.prompt.md). Copyright stays with the author.

Plan: model‑mapped task list TL;DR — Implement an optional Preact presentation plugin + Node service (MongoDB + socket.io) that adds multi-presentation storage, team-based ACLs, owner one‑time auth links, realtime cursors/diffs, bounded auto-history (N=50) + named snapshots, thumbnails (separate files), and client export (PNG + notes). Split the work into clear implementation steps and assign each step to an appropriate AI model tier (mini → low-cost for routine code, medium → design/complex flows, large → architecture/CRDT/security). This lets you run inexpensive agents for most work while reserving stronger models for hard problems. GPT 5 mini

1. Scaffold client + server: package.json, vite.config.ts, src/entry.tsx, server/src/index.ts, basic NPM scripts.
2. Create UI skeletons: src/presentation/plugin.tsx, ToolbarButton.tsx, PresenterView.tsx shells with ARIA.
3. Implement slide model helpers: src/presentation/slideModel.ts — getSlidesFromFile(), serializeSlide(), deserializeSlide().
4. Client export: src/presentation/export.ts using html2canvas to produce PNGs.
5. GridFS helper (basic): server/src/storage/thumbnail.ts to store/retrieve thumbnail files.
6. Simple unit tests: tests/slideModel.test.ts and test harness.
7. Docker/dev compose: minimal docker-compose.yml (MongoDB + optional Redis).

Claude Sonnet 4.5

1. Design and implement REST APIs + DB models: server/src/api/presentations.ts, auth.ts, teams.ts and models (user.ts, team.ts, presentation.ts, slide.ts, version.ts, snapshot.ts, sharelink.ts, authToken.ts).
2. Auth flows: owner one‑time token generation/accept flow, session issuance, email-as-username enforcement.
3. Team management: create-on-demand teams, add/remove member by username, user-search behavior and invite fallback.
4. Permissions middleware: server/src/middleware/perm.ts enforcing visibility/ACLs and sharelink logic.
5. Realtime server + client implementation: server/src/ws.ts (socket.io), src/presentation/rtc.ts client wrapper — protocol joinRoom, presence, diff, slideChange, snapshot.
6. Versioning job & snapshots: server/src/jobs/pruneVersions.ts, snapshot APIs, history UI src/presentation/history.tsx.
7. Presenter UX & pointer sync: implement pointer broadcast, smoothing, and thumbnails hooks.
8. Integration tests: tests/ws.integration.ts multi-client scenario.

Claude Hainku 4.5

1. Security & threat model: token lifecycle, revocation, CSRF/XSS/serialization hardening for scene JSON, sensitive endpoints review.
2. Scaling & architecture: multi-instance socket.io with Redis adapter, sticky session guidance, DB indexing and sharding considerations, autoscaling guidance.
3. CRDT migration plan: concrete migration path from server-authoritative diffs → yjs (data model, storage, dual-mode runbook).
4. Performance & storage strategy: audit log retention policy design, snapshot/thumbnail storage costs and pruning strategy, GridFS vs S3 decision with migration plan.
5. Compliance & privacy: PII handling rules (email-as-username), GDPR/retention implications, secure default configurations.
6. Final security review before production release and detailed mitigation steps.

Rapter AI mini (Preview)

1. Code quality & maintainability: set up linting, formatting, and commit hooks;
2. Developer experience: comprehensive README, contribution guidelines, API documentation, and onboarding guide.
3. Monitoring & observability: integrate logging, error tracking, and performance monitoring tools; define
4. Give ideas to test the code