CLAUDE.md@evals/trip-planner/iteration-1/new-trip-scaffolding/with_skill/outputs · git:20260609.5232324 · 2026-06-09 · sha256 7f6c6127722af440
CLAUDE.md@evals/trip-planner/iteration-1/new-trip-scaffolding/with_skill/outputs git:20260609.5232324A
Immutable. This exact content is served forever at /api/v1/blob/7f6c6127722af440.
# Memory
## Me
Tanisha & Alex — planning a 10-day trip to Japan in October 2026. Deep exploration of 2-3 cities. Food-focused, temples, gardens. Both first-timers to Asia. Budget: $8,000 USD. Alex has an old knee injury (stairs manageable but cumulative after long activity days) and is nervous about language barrier.
## People
| Who | Role | Notes |
|-----|------|-------|
| **Tanisha** | Primary planner, vegetarian | Detail-oriented, organized, experienced with logistics and research. First time in Asia. |
| **Alex** | Travel companion, omnivore | First time in Asia, nervous about language barrier. Knee constraint (stairs cumulative). Food-focused. |
→ Full profiles: `memory/people/tanisha.md` and `memory/people/alex.md`
## Trip Planning Status (Apr 1, 2026)
- **Current Phase:** ✅ **Phase 1 (Define)** — Complete. Project structure, traveler profiles, conventions established.
- **Next Phase:** **Phase 2 (Route)** — Research 2-3 Japan city options, create decision-support materials if needed, lock route.
- **Last session:** `memory/sessions/session-log-apr01-a.md` — Phase 1 scaffolding complete. Ready for city research.
### Trip Snapshot
| Item | Value |
|------|-------|
| **Dates** | Oct 10–19, 2026 (10 days, placeholder — confirm with Alex) |
| **Travelers** | Tanisha (vegetarian) + Alex (omnivore, knee + language anxiety) |
| **Cities** | 2-3 TBD (Phase 2) — likely mix of Kyoto, Tokyo, Osaka, Kanazawa, or Hiroshima |
| **Budget** | $8,000 USD total (~$4,000/person) |
| **Flights** | 85,000 United miles + cash if needed |
| **Trip Style** | Depth over breadth, food-focused, temples, gardens, minimal tourism |
### Budget Estimate
- Flights: $2,000–2,500 USD
- Accommodation (9 nights): $1,800–2,400 USD
- Activities: $600–900 USD
- Restaurants: $2,200–2,800 USD (food-focused; Japan affordable)
- Local transport: $300–400 USD
- Buffer: $500–800 USD
- **Total: ~$8,000 USD**
Booked so far: **$0** (Phase 1 only)
## Preferences
**Tanisha:**
- Vegetarian (strict — no meat, poultry, fish)
- Interested in: food, temples, gardens, authentic experiences
- Travel style: deep exploration, research-driven, detail-oriented
- Booking preference: unknown — ask
**Alex:**
- Omnivore (eats everything)
- Interested in: food (exploring Japanese cuisine), temples, gardens
- Knee consideration: stairs manageable on flat days, cumulative after activity. Plan around it naturally (location, pacing).
- Language anxiety: first time in Asia, nervous. Needs clear, visual guides and reassurance.
- Booking preference: unknown — ask
**Trip-wide:**
- Depth over breadth: 2-3 cities, 3+ nights each
- Food focus: both travelers prioritize eating well
- Authentic experiences: less tourist-heavy, more neighborhood-oriented
- Accessible (for Alex's knee): prefer elevators, lower floors, activities with rest points
- No hyper-planning: keep schedule flexible, allow for discovery
## Conventions (Project-Wide)
| Item | Value |
|------|-------|
| **Home currency** | USD (assume, confirm if CAD preferred) |
| **Temperature units** | Fahrenheit (assume, confirm if metric preferred) |
| **Distance units** | Miles (assume, confirm if km preferred) |
| **Dietary strict** | Tanisha: vegetarian (no meat/poultry/fish). Alex: omnivore (no restrictions). |
| **Loyalty programs** | United: 85,000 miles available. Other programs: TBD. |
| **Communication** | Tanisha: detail-oriented, data-driven. Alex: appreciates visual clarity, reassurance. |
**Booking philosophy:** NOT YET ESTABLISHED — ask Tanisha and Alex early Phase 2. Options:
- **Flexibility-first:** Refundable flights, hotels with free cancellation, premium options for comfort. Higher cost.
- **Cost-optimized:** Non-refundable fares, locked-in bookings, seeking best value. Lower flexibility.
- **Balanced:** Mix of flexibility and value.
This choice shapes every booking decision. Ask early.
## Glossary
Key terms and abbreviations. → Full glossary: `memory/glossary.md`
| Term | Meaning |
|------|---------|
| **Phase 1–8** | Planning phases from Define (currently here) through Final Check |
| **Route** | The cities, order, and night allocation for the trip |
| **CPP** | Cents per point — value of loyalty mile redemptions |
| **JR** | Japan Railways — major train operator |
| **Suica/Pasmo** | Prepaid transit cards for local transport |
| **Knee** | Alex's soccer injury constraint — plan for natural rest points |
## Open Questions
1. **Trip dates:** Confirm Oct 10–19 with Alex (or adjust if needed)
2. **Booking philosophy:** Flexibility vs. cost vs. balance? Shapes all future decisions.
3. **Loyalty programs:** Any others besides United 85,000 miles?
4. **Tanisha/Alex location preferences:** Any cities they've heard of or want to avoid?
5. **Hotel preferences:** Modern, traditional, mix?
6. **Dining style:** Casual, fine dining, mix?
## File Map
### Root — Core working docs
| Path | What |
|------|------|
| `CLAUDE.md` | This file — agent working memory |
| `TASKS.md` | Active task tracker — **start here each session** |
### memory/ — Core reference (read often)
| Path | When to read | What |
|------|-------------|------|
| `memory/itinerary.md` | **Most tasks** | Day-by-day schedule, accommodation, transport (once finalized) |
| `memory/projects/japan-trip.md` | **Trip structure questions** | Strategy, budget, constraints, open questions |
| `memory/people/tanisha.md` | **Activity/preference decisions** | Tanisha's profile: interests, travel style, dietary needs |
| `memory/people/alex.md` | **Activity/comfort/language decisions** | Alex's profile: knee constraint, language anxiety, interests, dietary needs |
| `memory/glossary.md` | **Unfamiliar abbreviations** | Trip terms, planning phases, Japan-specific shorthand |
### memory/sessions/ — Cross-session handoff
| Path | What |
|------|------|
| `memory/sessions/session-log-apr01-a.md` | **Latest (Apr 1)** — Phase 1 scaffolding complete. Ready for Phase 2 city research. |
### memory/research/ — Research files (created as needed)
| Path | What |
|------|------|
| (Not yet created) | Phase 2+: City comparisons, flight research, hotel shortlists, restaurant research |
### memory/guides/ — Activity/restaurant guides (created Phase 5-6)
| Path | What |
|------|------|
| (Not yet created) | Phase 5+: Per-city activity guides. Phase 6: Per-city restaurant guides. |
### memory/plans/ — Implementation plans (created as needed)
| Path | What |
|------|------|
| (Not yet created) | Phase 7+: Implementation plans for deliverables (daily cards, packing lists, etc.) |
### outputs/ — Traveler-facing materials
| Path | What |
|------|------|
| `outputs/tickets/` | Downloaded e-tickets and vouchers (created Phase 5+) |
| `outputs/receipts/` | Purchase receipts (created Phase 3+) |
| (Not yet created) | Phase 7: Itinerary guide, daily cards, packing lists, language sheets |
### archive/ — Superseded options
| Path | What |
|------|------|
| (Not yet created) | Rejected city options, old routes, cancelled bookings |
## Agent Workflow Notes
### Session Protocol
**At session start:**
1. Read CLAUDE.md (this file) — current phase, status, open questions
2. Read TASKS.md — what's active, next, blocked
3. Read the latest session log (linked above)
4. Summarize briefly (10-15 lines): current phase, active tasks, suggested focus
5. Ask user if they want to go with that or redirect
**At session end:**
1. Write a session log to `memory/sessions/session-log-{date}-{letter}.md`
2. Update CLAUDE.md with new phase (if changed), last session pointer, open questions
3. Reconcile TASKS.md: move completed tasks to Done section with strikethrough + date
### Session Log Format
Each session log includes:
- **Summary:** 2-3 sentences
- **Changes Made:** Bullet list of files created/modified
- **Key Decisions:** What was decided and why
- **Open Items:** Anything unresolved
- **Next Session:** What to focus on next
### General Principles
- **Check TASKS.md first** at session start — it's the task priority list
- **On disk is source of truth** — CLAUDE.md, TASKS.md, memory/ files are the real project state
- **Document decisions in project files** — not in session logs alone
- **Phases are natural progression, not rigid** — they can overlap
- **Research → Guide pipeline:** Research files are broad comparisons. Guides are curated verdicts created *after* user approval. Never create a guide before user input.
---
**Session initialized:** Apr 1, 2026. Ready for Phase 2 (Route) when user is ready.