content / skill

minto-pyramid

An Agent Skill for writing and restructuring documents in the Pyramid Principle: answer first, then grouped reasons, then evidence.

At a glance

Purpose
An Agent Skill for writing and restructuring documents in the Pyramid Principle: answer first, then grouped reasons, then evidence.
Runs in
Chat app with skills, Local agent with shell
Operating systems
Not stated
Needs
Runtime: Not stated; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data
Reads files: Not stated; Writes files: Not stated; Sends data to other services: Not stated
Tool type
Skill
Delivery
From GitHub
Version
Latest on the main branch
Access
Code on GitHub
Compatibility
According to the README, it works in Claude Code, Claude Desktop and anything else that reads Agent Skills. It shows install commands for a Bash-style shell and for Windows PowerShell, and says to restart your session afterwards. For models, it reports that the restraint behavior holds on Opus, Sonnet, GPT-5 and Gemini, and says it does not hold on Haiku.
External costs
The repository is MIT-licensed and states no pricing. Your coding agent or model provider may charge for usage.
Permissions
The skill acts through your coding agent with the access you give it. The README shows cloning or copying SKILL.md into ~/.claude/skills/minto-pyramid, with a matching Windows user-profile path, and gives an alternative of asking the agent to install from the repository address. The README describes the skill as one file without dependencies. The README says it offers a rewrite and waits unless the user explicitly requests the skill or asks the agent to write the document. The README does not describe additional file, shell or network access by the skill.
Usage terms
MIT License, as stated in the repository. See its LICENSE file for the exact terms.

Code on GitHub

millwright-labs/minto-pyramid-skill

MIT

80 stars7 forksLast pushed 16 Aug 2026 UTC

Agent Skill: make Claude write in Barbara Minto's Pyramid Principle - answer first, grouped reasons, evidence under each.

Topics: agent-skills business-writing claude-code claude-skills communication executive-communication minto pyramid-principle

Public GitHub data fetched · 0 days old

View on GitHub ↗

What it does

minto-pyramid is an Agent Skill for writing and restructuring decision documents. The README says it instructs a coding agent to put the answer first, followed by reasons and evidence, and apply Minto's grouping rules. The README says each operation includes an instruction to invent nothing.

The README says the skill offers a rewrite and waits when it loads on a shared draft, except when the user explicitly requests the skill or asks the agent to write the document. The README says it declines documents such as timelines, runbooks and thank-you notes when their structure does not suit a decision document.

The README names restructure, buried-lede test, reason audit, so-what pass and email version as operations. The README says users can request an operation by name or supply a draft for restructuring. The repository's top level includes SKILL.md, README.md, LICENSE and evals/.

Who it is for

People who use a coding agent to write or restructure decision documents, such as recommendation emails.

Example tasks

Use the minto-pyramid skill on a draft you shared.
Restructure a draft so the conclusion comes first.
Run the buried-lede test on a draft.
Run a reason audit on a recommendation.
Ask for the email version of a document.

Getting started

The README clones the repository into a minto-pyramid folder under your Claude skills folder, which is ~/.claude/skills in a Bash-style shell, with a matching PowerShell command for Windows. Alternatively, hand the repository address to your agent and ask it to install the skill, or download SKILL.md and place it at ~/.claude/skills/minto-pyramid/SKILL.md. Restart your session afterwards. The README has the exact commands.

Limitations

According to the README, the skill is deliberately narrow. The README says the skill declines incident timelines, runbooks, tutorials and anything where chronology is the content. The README reports that its restraint behavior, offering instead of rewriting and declining documents it should not touch, holds on Opus, Sonnet, GPT-5 and Gemini. It says the behavior does not hold on Haiku, which restructured an incident timeline into a postmortem template, invented figures and, on one run, matched a different installed skill. The README advises using a frontier model, or checking the output when you do not. It says one comparison run with the skill sharpened a fact beyond what the source said, which is why it now writes "invent nothing" into every operation. It expects the gain to be smallest on a short email in a frontier model that was already going to lead with the ask. The repository can change at any time; check its README before you install.

Data practices

The README does not describe data handling. What is sent where depends on your coding agent and model provider.

Community experience

Not rated yet

Sign in to share your experience.

Comments 0

No comments yet.

Sign in to comment

Report a listing concern