Date Night Skill (Published)

PII note: user_email, user_phone, first_name, and last_name are used to auto-fill reservation and ticket forms. They are stored locally in config.json and never transmitted except to the booking site during form submission.

First Run Onboarding

Check at every skill invocation:

CONFIG=~/.openclaw/skills/date-night/config.json
if [ ! -f "$CONFIG" ]; then
  echo "ONBOARDING_NEEDED"
fi

If ONBOARDING_NEEDED: run the onboarding flow before doing anything else.

Onboarding Flow

Ask questions one at a time, conversationally. This is a date night skill — make it feel warm and fun, not like filling out a DMV form.

Opening:

"Hey! Looks like it's your first time using the Date Night skill — exciting. Let me grab a few quick things so I can make it feel personal. Won't take long."

Question sequence (ask naturally, wait for each answer):

1. Name:

   "First — what's your name?" — Collect first name and last name (needed for reservation forms).

2. Email & Phone:

   "What email and phone number should I use for reservations? These go directly into booking forms — I store them locally and nowhere else."

3. Partner:

   "Planning these with a partner, or flying solo? (Solo date nights are completely valid.)" — If partner: "What's their name?"

4. Partner notification:

   "What's the best way to reach {partner}? Like, if I wanted to send them a heads-up about a reservation — iMessage, Telegram, Signal, Discord, something else?" — Skip if solo.

5. Dietary / lifestyle:

   "Any food preferences or restrictions I should know about? Things like no alcohol, vegetarian, gluten-free, shellfish allergy — or just 'none, we eat everything'?"

6. Kids:

   "Do you have kids at home? (I ask because I'll remind you about childcare after every booking.)" — If yes: "How many, and roughly what ages?"

7. Location:

   "What city or zip code are you in? I'll use this for finding nearby restaurants, theaters, and venues."

8. Theater preference (optional):

   "Any favorite movie theater chain, or a go-to theater near you? I can default to that when searching showtimes. (Skip if you don't have one.)"

9. Babysitter rate (optional, only if has_children):

   "What's your babysitter rate — roughly? I use it for budget estimates. Default is $18/hr if you're not sure."

10. Calendar tool:

    "Last one — how do you manage your calendar? I can add events automatically. Options: gog (Google), gcal, ical, or tell me what you use."

Closing:

"Perfect — you're all set! 🎉 Just say 'plan a date night,' 'find us a restaurant,' or 'get us tickets' anytime. I've got the rest."

Save Config

After collecting all answers, write the file:

mkdir -p ~/.openclaw/skills/date-night
cat > ~/.openclaw/skills/date-night/config.json << 'EOF'
{
  "name": "{name}",
  "first_name": "{first_name}",
  "last_name": "{last_name}",
  "user_email": "{email}",
  "user_phone": "{phone_digits}",
  "partner": "{partner_or_null}",
  "notify_channel": "{channel}",
  "dietary": ["{pref1}", "{pref2}"],
  "has_children": {true|false},
  "children_count": {N},
  "children_ages": "{ages_or_null}",
  "location": "{City, ST}",
  "zip": "{zip}",
  "preferred_theater": "{theater_or_null}",
  "babysitter_rate": {rate},
  "calendar_tool": "{tool}",
  "onboarded_at": "{ISO_TIMESTAMP}"
}
EOF

Applying Config Throughout the Skill

After loading config, substitute into all references:

Dietary preference enforcement: If dietary includes "no alcohol": never highlight wine lists, cocktail programs, bar scenes, or alcohol features when recommending restaurants or events. Focus on food, ambiance, service.

Reconfigure

If user says "update my date night preferences", "reconfigure date night", or "change my date night settings":

# Back up existing config
cp ~/.openclaw/skills/date-night/config.json \
   ~/.openclaw/skills/date-night/config.backup.json 2>/dev/null || true

Then re-run the onboarding flow with current values shown as defaults:

"Sure — let's update your preferences. I'll show what you have now and you can change anything. Hit Enter to keep the current value."

After re-running, overwrite config.json. Confirm:

"Updated! Changes are live for the next date night."

Requirements

External Binaries (must be installed separately)

PATH setup: export PATH="$HOME/.npm-global/bin:$PATH"

Environment Variables

OpenClaw Built-in Capabilities Used

These are standard OpenClaw agent tools — not external dependencies. They require no separate installation; they are available to any OpenClaw agent with the appropriate tool policies enabled.

No additional credentials or API tokens are needed for these built-in tools — they use whatever channels and connectors the user has already configured in their OpenClaw instance.

Sensitive Access Declaration

SMS verification codes: OpenTable and Resy send 6-digit verification codes via SMS during booking. This skill retrieves them by reading the most recent 1–2 messages from the specific sender (e.g., short code 22395 for OpenTable). It does NOT perform broad inbox scans. Specific patterns used:

# OpenTable: read last message from known short code only
imsg history --chat-id {OPENTABLE_CHAT_ID} --limit 1 | grep -oE '[0-9]{6}'
# Resy: read last 10 messages, filter for "resy" or "verification" only
imsg history --limit 10 | grep -i "resy\|verification\|code"

Gmail (modify/cancel only): When the user explicitly asks to modify or cancel a reservation and doesn't have the confirmation number, the skill searches Gmail with a targeted query — e.g., from:opentable reservation confirmed — limited to 5 results. This is never triggered during normal booking flows.

Partner notifications: Notifications are always drafted and shown to the user for approval before sending. The skill never sends messages autonomously. The messaging channel (iMessage/Telegram/Discord/Signal) is whatever the user configures in their OpenClaw instance — no additional credentials are stored by this skill.

Auth state persistence: The Resy flow optionally saves browser session state to ~/.openclaw/skills/date-night/state/resy-auth.json to avoid re-login. This file contains session cookies. To clear: rm -rf ~/.openclaw/skills/date-night/state/. The skill never asks for or stores site passwords — it uses interactive browser login and saves only the resulting session cookies with user consent.

Data Persistence

To purge all skill data: rm -rf ~/.openclaw/skills/date-night/

Pre-Flight Check

# Verify playwright-cli is available
export PATH="$HOME/.npm-global/bin:$PATH"
playwright-cli --version || echo "INSTALL: npm install -g @playwright/cli@latest"

References

Quick Flow: Dinner Reservation

export PATH="$HOME/.npm-global/bin:$PATH"
CONFIG=$(cat ~/.openclaw/skills/date-night/config.json)

1. Find restaurant — see search-restaurants.md
2. Open restaurant page:

   playwright-cli open "https://www.opentable.com/r/{slug}?covers={N}&dateTime={YYYY-MM-DDTHH:MM}" --headed
   

3. Snapshot → select date → select time
4. Fill guest details (name, phone, email from user's MEMORY.md or config)
5. Submit → handle phone verification (see sms-codes.md)
6. Confirm → post-booking actions (below)

Full flow details: opentable-flow.md | resy-flow.md

Quick Flow: Movie Tickets

1. Start at preferred theater (from config) or search via Fandango
2. Select movie, showtime, seats
3. ⚠️ STOP at payment — show all-in price, confirm with user before purchase
4. Post-purchase actions (below)

Full details: movie-booking.md | search-movies.md

Quick Flow: Event Tickets

Start with SeatGeek (best price discovery + Deal Score):

playwright-cli open "https://seatgeek.com/search?q={event}&location={config.location}" --headed

Then cross-check Ticketmaster (official) and StubHub (resale). ⚠️ Ticketmaster has aggressive bot protection — see event-tickets.md. ⚠️ STOP at payment — confirm all-in price with user before purchase.

Full Date Night Flow

When user says "plan a date night around [event/movie/idea]":

1. LOAD config.json silently
2. FIND the event/movie → SeatGeek + Ticketmaster + StubHub price comparison
3. PRESENT options + prices to user → get approval to proceed
4. BOOK tickets (with explicit user confirmation at payment)
5. NOTE venue and show time
6. SEARCH restaurants within 1 mile of venue, open before the show
7. SUGGEST 2-3 dinner options with ratings and OpenTable availability
8. BOOK dinner (with approval) at time that ends ~1 hr before show
9. CHECK weather for that evening → include in summary
10. CALCULATE timeline: leave home → dinner → venue → showtime
11. ADD both events to calendar ({config.calendar_tool})
12. DRAFT partner notification → show draft → send with approval
13. IF has_children: PROMPT about childcare — every time, no exceptions
14. OFFER dessert spot near venue (optional extension)
15. PRESENT budget estimate: dinner + tickets + babysitter total

Example timeline for 7:30 PM show: