Luma Events

Fetch upcoming events from Luma (lu.ma) for any city. Use when the user asks about tech events, startup meetups, networking events, conferences, or things happening in cities like Bangalore, Mumbai, Delhi, San Francisco, New York, etc.

Security Scan

The skill's code and network behavior match its stated purpose (scraping lu.ma pages), but SKILL.md directs persistent saving/merging of fetched events to ~/clawd/memory while the included script does not implement that persistence—this mismatch and the implicit persistent storage of user activity warrant caution.

Purpose & Capability

The name/description (fetch Luma events) align with the included Python script: it fetches lu.ma/<city>, extracts __NEXT_DATA__ JSON, filters and formats events. Declared dependencies (stdlib only) and package.json metadata are consistent with the stated purpose. Note: SKILL.md describes event persistence to a Clawd memory file as a feature, but the provided script does not implement writing or merging to ~/clawd/memory/luma-events.json — an implementation mismatch.

Instruction Scope

SKILL.md explicitly instructs the agent to 'Always save fetched events to ~/clawd/memory/luma-events.json' and to merge/retain 60 days of events. The runtime Python script performs network fetches and prints or emits JSON, but it contains no file-write or merge logic. This creates unclear runtime responsibility (will the agent itself write to disk?) and grants the skill the ability to persist user browsing/activity data if the agent follows SKILL.md. Persisting user activity is privacy-sensitive and should be explicit and consented to.

Install Mechanism

No install spec or downloads — this is instruction + script only and uses Python stdlib. No external packages or remote installers are pulled, which reduces install risk.

Credentials

The skill requests no environment variables or credentials and only contacts the public lu.ma domain, which is proportionate for event scraping. The only unusual permission is implicit filesystem persistence (writing to ~/clawd/memory), which is not declared as an environment/config requirement but is requested in documentation; that should be made explicit to the user.

Persistence & Privilege

The skill does not set always:true and does not request elevated system privileges. However, SKILL.md asks for always-saving fetched events to a user-local memory path, which grants persistent storage of the user's event queries and interests. Because the saved-file behavior is not implemented in the script, it's unclear whether persistence would occur automatically; this ambiguity increases the risk surface (privacy/long-term data retention).

What to consider before installing

The skill appears to do what it advertises (scrapes lu.ma public pages) and uses only Python stdlib, which is reasonable. However: (1) SKILL.md instructs the agent to always save fetched events to ~/clawd/memory/luma-events.json and to merge/retain them for 60 days, but the included script does not write or merge that file — ask the publisher which component (agent vs script) is responsible for persistence before installing. (2) If you allow automatic saving, be aware this will record the cities/events you query (privacy concern). Require explicit consent and review/controls for that memory file. (3) If you prefer, run the script manually in a sandbox first to confirm output, or modify the script to implement safe, explicit persistence (prompt before write, use a configurable path, or disable auto-save). (4) The network access is only to lu.ma; ensure you are comfortable with public web scraping and respect rate limits. If you need the skill to be non-persistent, remove or disable the 'save to ~/clawd/memory' instruction before granting the skill write permission.

Luma Events Skill

Fetch structured event data from Luma (lu.ma) without authentication. Luma is a popular platform for tech meetups, startup events, conferences, and community gatherings.

How It Works

Luma is a Next.js SSR app. All event data is embedded in the HTML as JSON inside a <script id="__NEXT_DATA__"> tag. The Python script extracts this data - no API key needed.

Quick Start

python3 scripts/fetch_events.py bengaluru mumbai --days 14

Usage

python3 scripts/fetch_events.py <city> [cities...] [--days N] [--max N] [--json]

Parameters

• city: City slug (bengaluru, mumbai, delhi, san-francisco, new-york, london, etc.)
• --days N: Only show events within N days (default: 30)
• --max N: Maximum events per city (default: 20)
• --json: Output raw JSON instead of formatted text

Popular City Slugs

• India: bengaluru, mumbai, delhi, hyderabad, pune
• USA: san-francisco, new-york, austin, seattle, boston
• Global: london, singapore, dubai, toronto, sydney

Output Format

Human-readable (default)

============================================================
📍 BENGALURU — 5 events
============================================================

🎯 AI Engineers Day with OpenAI
📍 Whitefield, Bengaluru
📅 Jan 31, 2026 10:30 AM IST
👥 OpenAI, Google AI
👤 1411 going
🎫 Available (150 spots)
🔗 https://lu.ma/57tarlkp

🎯 Startup Fundraising Masterclass
📍 Koramangala, Bengaluru
📅 Feb 02, 2026 06:00 PM IST
🟢 Free (50 spots)
🔗 https://lu.ma/startup-funding

JSON output (--json)

[
  {
    "city": "bengaluru",
    "count": 5,
    "events": [
      {
        "event": {
          "name": "AI Engineers Day",
          "start_at": "2026-01-31T05:00:00.000Z",
          "end_at": "2026-01-31T12:30:00.000Z",
          "url": "57tarlkp",
          "geo_address_info": {
            "city": "Bengaluru",
            "address": "Whitefield",
            "full_address": "..."
          }
        },
        "hosts": [{"name": "OpenAI", "linkedin_handle": "/company/openai"}],
        "guest_count": 1411,
        "ticket_info": {
          "is_free": false,
          "is_sold_out": false,
          "spots_remaining": 150
        }
      }
    ]
  }
]

Event Persistence

Always save fetched events to ~/clawd/memory/luma-events.json for future reference.

This allows you to:

• Answer questions about events without repeated fetches
• Track which events the user is interested in
• Compare events across cities
• Build context about upcoming plans

When to save:

• After fetching events for any city
• Merge with existing data (by event URL)
• Keep events for next 60 days only
• Add lastFetched timestamp

Format:

[
  {
    "city": "bengaluru",
    "name": "AI Engineers Day",
    "start": "2026-01-31T05:00:00.000Z",
    "end": "2026-01-31T12:30:00.000Z",
    "url": "https://lu.ma/57tarlkp",
    "venue": "Whitefield, Bengaluru",
    "hosts": ["OpenAI", "Google AI"],
    "guestCount": 1411,
    "ticketStatus": "available",
    "spotsRemaining": 150,
    "isFree": false,
    "lastFetched": "2026-01-29T12:54:00Z"
  }
]

Common Use Cases

Find tech events this week

python3 scripts/fetch_events.py bengaluru --days 7

Check multiple cities for AI events

python3 scripts/fetch_events.py bengaluru mumbai san-francisco --days 14 --json | jq '.[] | .events[] | select(.event.name | contains("AI"))'

Get next 5 events in a city

python3 scripts/fetch_events.py new-york --max 5

Example Queries

User: "What tech events are happening in Bangalore this weekend?" → Fetch Bengaluru events for next 7 days, save to memory

User: "Any AI meetups in Mumbai next month?" → Fetch Mumbai events for next 30 days, filter for AI-related, save to memory

User: "Compare startup events in SF vs NYC" → Fetch both cities, compare, save both to memory

Notes

• No authentication: Luma event pages are public
• City slugs: Use lowercase, hyphenated slugs (san-francisco, not San Francisco)
• Rate limiting: Respectful fetching only (don't hammer the servers)
• Data freshness: Events are live data from the HTML, always current
• Timezone: Times are in the event's local timezone (extracted from start_at)

Troubleshooting

"Could not find NEXT_DATA" → Luma changed their HTML structure, script needs updating

"Unexpected data structure" → The JSON path changed, check the latest HTML

No events returned → City slug might be wrong, or no upcoming events for that city

Timeout errors → Network issue, retry or check internet connection

Dependencies

• Python 3.6+ (stdlib only - no external packages needed)
• urllib, json, re, argparse, datetime (all built-in)

Changelog

v1.0.0 (2026-01-29)

• Initial release
• Support for multiple cities
• Human-readable and JSON output
• Date filtering (--days)
• Event limit per city (--max)
• Event persistence to memory file