signup-monitor

name	signup-monitor
description	Cron-triggered: query PostHog for new signups, research each by email domain via Exa, alert the CEO when someone interesting signs up. Not for manual use — runs automatically.
user-invocable	false

Signup Monitor

Automatic lead research on every new [YOUR_PRODUCT] signup. Runs as a cron job.

Workflow

Step 1: Load State

Read memory/signup-monitor.json. If missing, create with:

{"last_checked": "2026-03-28T00:00:00Z", "processed": []}

Step 2: Query New Signups

posthog__query-generate-hogql-from-question \
  question="Show User Signup events after TIMESTAMP with properties org_owner_email, org_owner_name, org_name, org_member_count, organisation_id, signup_source, referrer, acquisition_channel. Order by timestamp desc, limit 50"

Replace TIMESTAMP with last_checked value from state.

If zero results — update last_checked to now, save state, respond HEARTBEAT_OK.

Step 3: Filter & Deduplicate

For each signup:

1. Skip if org_owner_email is null/empty
2. Skip if email already in processed list
3. Extract domain from email (part after @)
4. Skip personal emails: gmail.com, yahoo.com, hotmail.com, outlook.com, mail.ru, yandex.ru, icloud.com, protonmail.com, live.com, aol.com
5. Remaining signups = candidates for research

If no candidates after filtering — update state, respond HEARTBEAT_OK.

Step 4: Research Each Candidate

For each candidate, run ONE Exa search on the email domain:

exa-search__web_search_advanced_exa \
  query="DOMAIN company" \
  category=company \
  numResults=5 \
  type=auto

From results, extract:

• Company name and what they do
• Approximate team size (if available)
• Industry / tech stack signals
• Funding stage (if available)

Step 5: Qualify

Score each candidate as HOT / WARM / SKIP:

HOT (alert immediately):

• Known tech company (>100 employees)
• Company matches your ICP profile from MEMORY.md (industry, tech stack, team type)
• Person title suggests decision-maker (VP, CTO, Director, Head of, Lead, Manager)
• Recent funding or hiring signals

WARM (log, include in daily recap):

• Business email but small/unknown company
• Relevant industry but no strong signals

SKIP (log only):

• Unresearchable domain
• No relevant signals found

Step 6: Alert

For each HOT candidate, send a chat alert:

Signup: [Name] ([email])
[Company Name] — [one-line description]
[Team size if known] | [Industry] | [Funding if known]
Signal: [why this is interesting — 1 line]
Source: [signup_source] / [referrer]

Reply "research [email]" for full dossier.

For WARM candidates, append a summary line at the end:

Also signed up: [N] warm leads ([domain1], [domain2]...)

Step 7: Update State

Write memory/signup-monitor.json:

{
  "last_checked": "CURRENT_TIMESTAMP",
  "processed": ["[e-mail]", "[e-mail]", ...],
  "stats": {
    "total_checked": N,
    "hot": N,
    "warm": N,
    "skip": N,
    "last_run": "CURRENT_TIMESTAMP"
  }
}

Keep processed list trimmed to last 500 entries to avoid bloat.

Log results via structured-log.

Notes

• Research costs 1 Exa call per candidate (~$0.01). At current volume (3/week) this is negligible.
• If volume grows past 20/day, batch research into subagents.
• When CEO replies "research [email]", use exa-company-research + exa-people-research + lead-enrichment for full dossier, then add to pipeline via pipeline-manager.

Related Skills

• posthog-analytics — PostHog query patterns and event catalog
• exa-company-research — deep company research
• exa-people-research — people research by name/role
• lead-enrichment — full contact profile (email, phone, LinkedIn)
• pipeline-manager — add qualified leads to pipeline
• signal-scanner — complementary signal detection