Revive Stalled LinkedIn DM Threads with an AI Agent — Your Inbox Is Pipeline

# Rainmaker — Loadout

```yaml
loadout: rainmaker
tier: apex                # Apex tier (or an active Apex trial)
mcp: https://apex.leadshark.io/mcp
```

**The job:** Your inbox is pipeline you already paid for. Rainmaker sweeps your DM threads for conversations that stalled while still warm, reads where each one actually left off, and drafts the pickup that continues the thread — not a cold 'just bumping this.' Booked meetings come from resurrected threads, not prospect #501.

## Before you run — ask the user

Only ask what you don't already know — check the conversation, your memory,
and the account itself first. Confirm anything you inferred in one line
instead of re-interviewing.

1. **What makes a thread worth reviving?** (ICP fit, buying intent, how stale is too stale)
2. **What's the goal of a pickup — book the call, get an answer, or just restart the thread?**
3. **Daily cap on revivals, and who's off-limits?** (customers, do-not-engage)

## Steps

1. **See** — call `list_recent_messages`: Sweep your recent DM threads — every conversation, including the ones that went quiet.
2. **See** — call `get_messages_with_person`: Read the full thread for each candidate — what was said, what was promised, where it stalled.
3. **Decide** — call `get_lead_activity`: Score the warmth: did they open what you sent, have they engaged since, is the intent still live?
4. **Decide** — call `manage_activity_limits`: Preflight: check the remaining DM budget — no pickup staged into a cap that's already spent.
5. **Decide** — call `list_actions`: Preflight: check the action log — anyone already touched by any Loadout gets skipped, never double-touched.
6. **Act** — call `send_message`: Draft the pickup that references the real last exchange and moves toward your goal — approval-gated.

## STOP — approval gate

Present everything you found and drafted BEFORE any write. Do not send,
schedule, publish, connect, or create automations until the user approves.
Edits beat re-drafts: apply their changes, don't start over.

## Never

- Write to LinkedIn without explicit approval this run.
- Invent facts, numbers, results, or people.
- Contact anyone on the do-not-engage list, or double-touch someone
  (check `list_actions` first).
- Retry around a refusal — report what was refused and why.

## Done means

- The user approved the drafts (aim for ≤2 edit rounds).
- Approved actions are queued or scheduled — verify via `list_actions`.
- Report back: what ran, what's staged, and where to monitor it.

## Why this stays account-safe

The warmest surface there is: only people who already talked to you. Rainmaker is read-heavy with a single gated write — it never invents context (every draft quotes the real thread), skips customers and do-not-engage, paces inside your daily budgets, and every pickup waits in the approval queue. Reviving a thread should feel like you remembered — because you approved every word.

Using Apex, sweep my recent LinkedIn DM threads for conversations that stalled while still warm. For each candidate, read the full thread and check their lead activity — did they open what I sent, have they engaged since? Rank by ICP fit and live intent, and for the best ones draft a pickup that references what we actually said last and moves toward {goal}. Skip customers and do-not-engage, respect my daily caps, and queue every message for my approval.