meeting-prep

What this skill does

# Daily Meeting Prep

## Config — read before starting

Read `../config/user.json` (resolves to `~/executive-assistant-skills/config/user.json`).
Extract and use throughout:

- `name`, `full_name` — user's name
- `primary_email`, `work_email` — Gmail accounts to check
- `whatsapp` — WhatsApp number for delivery
- `timezone` — IANA timezone (e.g. America/Argentina/Buenos_Aires)
- `slack_username` — Slack DM target
- `workspace` — absolute path to OpenClaw workspace (e.g. ~/.openclaw/workspace)

Do not proceed until you have these values.

## Debug Logging (MANDATORY)

Read `../config/DEBUG_LOGGING.md` for the full convention. Use `python3 {user.workspace}/scripts/skill_log.py meeting-prep <level> "<message>" ['<details>']` at every key step. Log BEFORE and AFTER every external call (gog, mcporter, Granola, web search). On any error, log the full command and stderr before continuing.

## Scope

- Timezone: {user.timezone}
- Calendars: primary of {user.primary_email} AND {user.work_email}
- Today's meetings only

**Timezone note:** Use explicit ART-bounded ISO8601 timestamps for calendar queries, NOT `--date`. Example: `gog calendar list primary --account <email> --from "2026-03-03T00:00:00-03:00" --to "2026-03-04T00:00:00-03:00" --json`. The `--date` flag uses UTC boundaries which misaligns with ART.

- **ALL meetings with attendees** — both external and internal
- Skip personal/solo events with no attendees (e.g. "Personal Trainer", "Golf", all-day reminders)

## Meeting types

### External meetings (attendees outside your email domains)

Full research brief (email context, Granola, LinkedIn, company research) — see below.

### Internal meetings (all attendees from your email domains: hypergrowthpartners.com, growth.li)

Lighter brief — no LinkedIn/company research needed, but still include:

- Attendee list
- Granola context from previous instances of this recurring meeting
- Recent email threads related to the meeting topic/agenda
- Any open action items from last time
- Format: same structure but skip LinkedIn/company sections

### Recurring collaborative meetings (e.g. podcasts, content sessions with external co-hosts)

These are external meetings — give them full briefs. Don't skip recurring meetings just because they're familiar.

## Error handling

- If `gog calendar` fails for one account: continue with the other account, note "⚠️ [account] calendar unavailable" in output.
- If Granola/Grain fails: continue without meeting history, note it per meeting.
- If WhatsApp delivery fails: attempt Slack delivery. If both fail, save to state file and report error.

## For each meeting

### 1. Event basics

Title, local time ({user.timezone}), attendees.

**RSVP status (MANDATORY):** For each attendee, check `responseStatus` from the calendar event:

- `accepted` → no flag needed
- `needsAction` → flag as "⚠️ hasn't responded"
- `declined` → flag as "❌ declined"
- `tentative` → flag as "❓ tentative"

If ANY non-organizer external attendee has NOT accepted (`needsAction`, `tentative`, or `declined`), add a visible line in the brief:
> ⚠️ *RSVP:* <name> hasn't accepted yet

This is informational — it doesn't mean they won't join, but it's useful to know ahead of time, especially for first calls or important meetings.

### 2. Email context (90-day lookback, with historical fallback)

Search Gmail both accounts for exchanges with attendees. For EACH attendee, search using these strategies in order:

1. **Email address** (primary — from calendar invite): `from:<email> OR to:<email>`
2. **Full name**: `"firstname lastname"`
3. **First name + "intro"**: `"intro firstname"` (catches informal intro subjects)
4. **First name + company**: `"firstname companyname"`

The attendee's email from the calendar invite is the most reliable identifier — always start there.

**Intro discovery** (after general email search):
5. Search for intro emails involving the attendee: `subject:intro <email>`, `subject:intro <firstname>`, `subject:introduction <email>`
6. Also check threads where a third party CC'd/introduced the attendee

**Recent email context** (after intro discovery):
7. Pull the most recent email threads with this attendee (by email address) to surface any recent updates, asks, or context leading into today's call

**Historical fallback** (if no results from 90-day search):
8. Run a broader search with NO date filter: `from:<email> OR to:<email>` — this catches long-standing relationships where the last email was months/years ago. If older threads exist, this is NOT a first call — note the relationship history.

- First call vs follow-up? Base this on ALL email history found (including historical), not just 90-day window
- **If first call: MUST include "who introduced + when" (date) if found in email; if not found, explicitly say "No intro trail found in email"**
- If follow-up: extract updates since last call
- **If email contains a concrete commercial trigger** (pricing, deliverables, scope, budget, urgency, timeline, decision-maker request), include it explicitly in the brief

### 3. Granola context

**Search by ATTENDEE, not by meeting title.** The same recurring meeting may have different titles week to week. Always search by the attendee's name or email to find all past meetings with them.

```bash
# Primary: search by attendee name
mcporter call granola.query_granola_meetings query="meetings with [attendee full name]"

# Fallback: search by company if attendee name yields no results
mcporter call granola.query_granola_meetings query="meetings with [company name]"
```

**Cross-check with `list_meetings`:** If the query results seem stale (oldest match is weeks old but you expect more recent), also run `list_meetings` for `last_week` or `this_week` and scan the participant lists for the attendee's email or name. This catches meetings where the title doesn't mention the attendee or company.

- **Recent (< 3 weeks):** Provide a richer summary (not one-liner): decisions made, key tensions, explicit action items, owners, and unresolved questions
- **Older (3+ weeks):** Broader context — relationship history, past decisions, recurring themes
- **No attendee match but company match exists:** Use company-level context and label it clearly as company-level
- **No results / first meeting:** Note that, provide email context instead
- Preserve citation links `[[N]](url)`
- Include a short explicit line: **"Why this meeting now"** based on prior action items or current email trigger
- **Exact name matching**: When attributing Granola results to an attendee, verify BOTH first AND last name match exactly. Different people can share a first name — never assume a match based on first name alone.
- **Auth failure:** If Granola returns an auth error, run `mcporter auth granola --reset` and retry once. If still failing, note "⚠️ Granola unavailable" and continue without it.
- **Empty summary:** If Granola returns a meeting record but with no/empty summary, note "Previous meeting found but no summary available" — don't silently skip it.

### 4. LinkedIn research

- Search: `"[attendee name] [company] LinkedIn"`
- Extract: current role, background, recent posts/activity

### 5. Company research

- Search: `"[company] recent news"`
- Search: `"[company] funding crunchbase"` (if startup/VC relevant)
- Extract: company stage, announcements, what they do

## Research rules

Read `{user.workspace}/style/MEETING_PREP_RULES.md` for additional research steps.

## Output format

Send via WhatsApp ({user.whatsapp}) AND Slack (DM to {user.slack_username}). **One message per meeting, chronological order, is mandatory.**

**Also send to Chief of Staff:** After sending all meeting briefs, upload the markdown brief file (`{user.workspace}/state/meeting-prep-YYYY-MM-DD.md`) to {user.chief_of_staff.name}'s Slack DM. Use the Slack API `files.upload` (or `files.uploadV2`):

```bash
curl -s -F file=@{user.workspace}/state/meeting-prep-YYYY-MM-DD.md \
  -F channels={user.chief_of_staff.slack_dm_chan