Call Followup Sweep logo

Call Followup Sweep

Organization
swan-gtm
call-followup-sweep

Use this skill when a day or a week of customer calls has piled up and the follow-through is owed but unwritten. Trigger on "review my calls this week", "draft my follow-ups", "what do I owe people", "what did I promise on these calls", "catch me up on yesterday's calls", "did anything fall through this week". Asks which window to sweep, reads every call in it, extracts the commitments — especially the ones you made — and returns one drafted email per account that earns one, the list of calls that earn none and why, and the promises still open from earlier sweeps. Produces drafts only; it never sends.

Overview

Publisherswan-gtm
Repositorygtm-skills
Skill namecall-followup-sweep
Stars
216
Forks
103
Bundled files
3
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by swan-gtm on GitHub. Read the source before you install it.

Installation

Install the Call Followup Sweep AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/swan-gtm/gtm-skills/tree/main/skills/boaz-descalo/call-followup-sweep
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/swan-gtm/gtm-skills.git /tmp/gtm-skills
mkdir -p .claude/skills
cp -r /tmp/gtm-skills/skills/boaz-descalo/call-followup-sweep .claude/skills/call-followup-sweep
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Call Followup Sweep in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Call Followup Sweep on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Call Followup Sweep is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Applies when a batch of customer calls has happened and the follow-through is owed. Produces a per-account draft for the calls that earn one, a no-email list with reasons for the calls that don't, and an aging report of commitments still open from prior sweeps. Nothing is ever sent.

The sweep

  1. Ask the window before anything else. Day or week — never guess, and never default to "all recent calls". The window changes the output: a day sweep is mostly fresh drafts while the call is still warm; a week sweep is mostly an aging and pattern report with fewer, denser drafts. If the answer is "week", confirm whether it means the trailing five business days or the calendar week — on a Wednesday those differ by half a book.
  2. Pull every call in the window from wherever calls are recorded — transcripts, recordings, or the user's own notes. Capture account, attendees, date, and duration for each. A scheduled call with no recording is a finding, not a gap: list it as unverified rather than dropping it silently.
  3. Extract commitments, not summary. Read references/commitment-extraction.md before the first call. Four things get pulled from every transcript — what you committed to, what they committed to, the question you didn't answer, and any date that was named out loud. The first category is the one that loses deals and the one reps skip.
  4. Triage which calls earn an email. Read references/triage-rubric.md. Most calls in a given week do not. The value of a sweep is subtraction — a run that drafts an email for every call has done nothing a calendar couldn't.
  5. Dedupe by account, not by call. Three calls with one account in one week is one email to that account, built from all three. Separate drafts to the same company arriving the same morning read as an unmanaged vendor.
  6. Draft to the call's shape. Read references/draft-patterns.md — the email a stalled renewal needs is not the email a first discovery needs. Cap every draft at 150 words and three visible action items; overflow goes to a named next call, not a longer email.
  7. Report the sweep, don't just dump drafts. Lead with commitments aged past three business days, then the drafts, then the no-email list with a one-line reason each, then any pattern that showed up across calls — the same objection landing three times in a week is worth more than any single draft.

What good looks like

  • The best operator extracts their own promises first. Reps instinctively mirror the customer's pain back at them, because that is what the transcript is loudest about — but the thing that actually rots a deal is the seller's unshipped "I'll get you the security doc by Thursday". A sweep that surfaces customer pain and misses two of the user's own open promises has failed at its only irreplaceable job.
  • The common mistake is the recap email. Both people were on the call; a narration of it is filler that buys nothing. The follow-up exists to date the commitments, put in writing the one thing that would otherwise be disputed later, and hand the champion something forwardable to people who weren't there. If a draft would still make sense with the account name swapped out, it is a template, not a follow-up.
  • Good output is auditable and mostly negative space: every action item carries a named owner and a date, the no-email list is typically longer than the draft list, aging commitments appear above fresh drafts, and the user can see at a glance which calls were read and which could not be.

Rules

  • MUST ask whether the window is a day or a week before reading anything.
  • MUST produce drafts only. NEVER send, schedule, or queue a message, and never write to the CRM as part of a sweep — the user reviews and sends by hand.
  • MUST attach a named owner and an explicit date to every action item, or state plainly that the call never produced one.
  • MUST list calls that were skipped or unreadable, with the reason, rather than returning a clean report over partial coverage.
  • NEVER invent a commitment, a date, or a number that was not said on the call; quote or flag the absence.
  • NEVER write in a register the user has not used — a draft that has to be rewritten for voice is worse than no draft.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Call Followup Sweep AI skill do?

Use this skill when a day or a week of customer calls has piled up and the follow-through is owed but unwritten. Trigger on "review my calls this week", "draft my follow-ups", "what do I owe people", "what did I promise on these calls", "catch me up on yesterday's calls", "did anything fall through this week". Asks which window to sweep, reads every call in it, extracts the commitments — especially the ones you made — and returns one drafted email per account that earns one, the list of calls that earn none and why, and the promises still open from earlier sweeps. Produces drafts only; it n...

Why use Call Followup Sweep on TypingMind?

Because you install it once and use it with any model. Call Followup Sweep is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Call Followup Sweep in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/swan-gtm/gtm-skills/tree/main/skills/boaz-descalo/call-followup-sweep. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Call Followup Sweep?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Call Followup Sweep?

As many as you like. As long as a model supports skills, you can use Call Followup Sweep with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Call Followup Sweep AI skill free?

Yes. It is published on GitHub by swan-gtm under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇