Deliverability audit
Cold emails land in spam or reply rates suddenly drop.
You get Ranked problems · Fix plan · Recovery timeline
Use this when a target meets the limits of safe sending. It answers whether the goal is reachable, what infrastructure it needs, and how long until it runs at full speed.
Use it when you need to know how many mailboxes, domains and LinkedIn accounts a meeting target requires, or how much a team can safely send.
“We need 30 meetings a month from outbound. How many mailboxes and LinkedIn accounts?”
What happens when you ask your agent with this skill installed. Every step below comes straight from the skill.
We need 30 meetings a month from outbound. How many mailboxes and LinkedIn accounts?
sending-capacity-plannerReady for your review. Nothing is sent or changed until you approve.
What good looks like
What you bring
What you get
Use this when a target meets the limits of safe sending. It answers whether the goal is reachable, what infrastructure it needs, and how long until it runs at full speed.
Meetings needed per month: M
÷ meeting rate from positive replies (default 50%) = positive replies
÷ positive reply rate per contact (default 1.5%) = contacts per month
× touches per contact (email steps) = emails per month
÷ working days (21) = emails per day
Example: 40 meetings ÷ 0.5 = 80 positive replies ÷ 0.015 = 5,333 contacts × 4 emails = 21,333 emails a month ÷ 21 = about 1,016 emails a day.
Mailboxes = daily emails ÷ 30 (a conservative per-mailbox cold volume). Domains = mailboxes ÷ 3. Add 20% spare. Example: 1,016 ÷ 30 = 34 mailboxes → 41 with spare, across about 14 domains.
Per account: about 15-25 invitations a day, so roughly 100 a week. If LinkedIn carries part of the load, divide contacts between channels and size accounts accordingly. Each account must belong to a real person on the team.
5,333 new contacts a month is about 64,000 a year. Compare with the SAM: if the ICP has 20,000 reachable contacts, the plan exhausts it in four months. That is the real bottleneck, and the fix is better conversion, not more mailboxes.
New mailboxes need 4-6 weeks to reach full volume. Show month by month capacity and the monthly cost of domains and mailboxes.
Usually one of: list size (SAM), reply rate (copy and targeting), meeting conversion (reply handling speed), or infrastructure. Raising positive reply rate from 1.5% to 3% halves every number above.
Goal: M meetings/month
Funnel: contacts N → positive replies N → meetings M (rates used, assumed or measured)
Email: N/day → N mailboxes, N domains (incl. spare)
LinkedIn: N accounts (if used)
List check: N contacts/month vs SAM N → months of runway
Ramp: month 1 X%, month 2 Y%, full from week 6
Bottleneck: <one thing to fix first>
Once the skill is installed, plain requests like these load it automatically. Swap in your own company, buyer and numbers.
“We need 30 meetings a month from outbound. How many mailboxes and LinkedIn accounts?”
“Here are our real conversion rates. Is 60 meetings a month realistic?”
“Plan our sending capacity for next quarter and tell me the bottleneck.”
Use this when a target meets the limits of safe sending. It answers whether the goal is reachable, what infrastructure it needs, and how long until it runs at full speed. You get: contacts and sends needed, mailboxes domains and accounts needed, ramp timeline, bottleneck.
Use it when you need to know how many mailboxes, domains and LinkedIn accounts a meeting target requires, or how much a team can safely send.
Run /plugin marketplace add prosyo-ai/gtm-skill and then /plugin install gtm-skills@prosyo to add the whole library, or run npx skills add prosyo-ai/gtm-skill –skill sending-capacity-planner for this skill alone. In Claude.ai, download the .zip from this page and upload it under Settings, Capabilities, Skills.
Have these ready: meeting goal, current conversion rates, channels and limits. If something is missing, the skill tells the agent to ask for it rather than guess.
Yes. The skill is a plain SKILL.md file with no vendor lock-in. Agents that load skills natively (Claude Code, Claude.ai, Codex, Cursor, Gemini CLI and others) pick it up automatically; anywhere else, paste the file into the conversation as instructions.
No. Every skill in the library drafts, researches and analyses, and waits for a person to approve anything that sends, schedules or changes records in bulk. Your inbox, CRM and accounts stay under your control.
Yes. It is free and MIT-licensed. Copy it, adapt it to your market and voice, and if you improve it, open a pull request so everyone gets the better version.
Paste this into Claude, Codex or any agent. It interviews you about your play, writes the skill in the house format, checks it and opens the pull request. Once it is merged, it shows up here with your name on it.
Fetch https://raw.githubusercontent.com/prosyo-ai/gtm-skill/main/AGENTS.md and follow its submission flow. Interview me about my GTM play (the real numbers, rubrics and examples), draft it in their format with reference pages, validate it and open the pull request.
The skill gives your agent the playbook. Penny, the Prosyo AI SDR, runs plays like this for you every day across LinkedIn and email.
Penny, the Prosyo AI SDR, finds your buyers, writes from what they actually posted, follows up on LinkedIn and email and lands every reply in one inbox with an answer ready. 7 days free, no card.