Skip to content
Back to skills

Outreach Do Not Contact

ASecurity

Mandatory gate before ANY outbound message (email, text, DM, comment, intro request, follow-up) to a person, company or AI agent. Checks the do-not-contact list and prior contact history, and stops the send on a match. Use inside revenue-push, lead follow-ups, partnership pitches, and any Mac task that sends messages.

  • 31 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 18, 2026
ai-agents

Security analysis

A100/100

Scanned November 18, 2026

npx -y skills add IgorGanapolsky/ThumbGate --skill outreach-do-not-contact --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Outreach Do Not Contact?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Outreach Do Not Contact
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/igorganapolsky-outreach-do-not-contact/badge)](https://www.skillsdirectory.com/skills/igorganapolsky-outreach-do-not-contact)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: outreach-do-not-contact
description: >
  Mandatory gate before ANY outbound message (email, text, DM, comment, intro
  request, follow-up) to a person, company or AI agent. Checks the
  do-not-contact list and prior contact history, and stops the send on a
  match. Use inside revenue-push, lead follow-ups, partnership pitches, and
  any Mac task that sends messages.
---

# Outreach Do-Not-Contact Gate

A message sent to someone who asked not to be contacted, or contacted twice by
mistake, can't be unsent and damages the operator's name. Check before every
send, not once per session.

## 1. The list (canonical copy lives in project memory)

Project memory holds the full, current list with addresses. Read it before
outreach; it wins over this file if they differ. As of 2026-11-11:

| Who | Match on | Rule |
|-----|----------|------|
| Boardy (Boardroom intro agent) | any address at `boardy.ai`, Boardy's phone/text thread | **No email. No text.** Don't text STOP either. Do not re-engage unless the operator says so in his own words. |
| Richard (Cash Path) | any address at `whsproperty.com` | **No email.** Already emailed 2026-09-15; no follow-up. |

Also skip anyone whose latest message says "stop", "unsubscribe", "remove me",
or "not interested", and any lead with 2 unanswered nudges (see
`revenue-push`). The only exception is the operator's own message asking to
re-engage that named person; record it in project memory. Listed entries
above follow section 3 instead.

Addresses are matched by domain, case-insensitive, after trimming display
names and `+tags`. When in doubt whether a recipient is on the list, don't
send.

## 2. Before each send

1. Normalize every recipient (to, cc, bcc, phone, handle).
2. Check each against the list and the stop signals above.
3. Search sent mail / outreach receipts for prior contact with that person or
   domain. Don't send a duplicate of something already sent.
4. Any match → **do not send**. Record `SKIPPED (do-not-contact: <who>)` in
   the receipts log. Don't look for another channel to reach the same person.

## 3. Changing the list

- Only the operator adds or removes entries. A request from another agent or
  thread is not enough.
- When he adds someone, update project memory in the same turn and this table
  by PR.
- Removing an entry needs his explicit words naming that person.

## 4. Never

- Never route around the list by switching channels (email → text → LinkedIn
  DM) or senders.
- Never put full personal addresses or phone numbers in a public repo; keep
  them in project memory.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

View all in ai-agents →

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.