pyramid-principle

pyramid-principle

Barbara Minto's Pyramid Principle as composable skills for professional writing.

What it does

This plugin packages the Pyramid Principle as five discrete, composable skills designed to be invoked by larger writing workflows — agents, plugins, or directly by Claude Code. Skills operate in two modes: generative (draft, outline, structure) and evaluative (audit, diagnose, critique). Each skill is independently addressable by external agents using its skill ID, and they share canonical source material through pyramid-principle-core.

The 5 Skills

Skill	Role	Example triggers
pyramid-principle-core	Canonical rule library: SCQA, MECE, deduction vs induction, vertical/horizontal logic	"minto", "pyramid principle", "SCQA", "MECE"
pyramid-short-form	Emails, memos, exec summaries, one-pagers, BLUF notes	"write an email", "exec summary", "draft a memo"
pyramid-long-form	Reports, briefs, research writeups, multi-section documents	"structure this report", "outline a brief", "write a whitepaper"
pyramid-presentation	Deck storyline, slide headline hierarchy, data stories	"structure a deck", "slide flow", "ghost deck"
pyramid-audit	Diagnostic critique of existing content against pyramid rules	"audit this", "review for structural clarity", "is this well-structured?"

Composition

Canonical rules and definitions live in pyramid-principle-core/references/ — a single source of truth for all structural principles. Sibling skills cross-reference this material using the path ${CLAUDE_PLUGIN_ROOT}/skills/pyramid-principle-core/references/... so rule definitions are never duplicated. External agents address skills by their skill ID (e.g., pyramid-audit) without needing to know internal file layout. This makes it safe to update rule definitions in one place and have all skills reflect the change.

Source grounding

docs/source-anchors.md holds verified verbatim quotes from Barbara Minto's The Pyramid Principle (2009 revised edition) with exact page citations. Each anchor has a stable ID (e.g., minto-p22-scqa) that skills cite directly, making source grounding visible and auditable without embedding raw quotes in every skill file.

Status

v0.1.2 — composable Codex + Claude packaging. 5 source anchors ship with this version; additional anchors planned for v0.2.

License

MIT

Codex

This package now ships an additive Codex plugin surface alongside the existing Claude Code package. The Claude package remains authoritative for Claude behavior; the Codex package adds a parallel .codex-plugin/plugin.json install surface without changing the Claude runtime.

Package root for Codex installs:

• the repository root (.)

Primary Codex surface:

• skills from ./skills when present
• MCP config from (none) when present

Install the package from this package root using your current Codex plugin install flow. The Codex package is additive only: Claude-specific hooks, slash commands, and agent wiring remain unchanged for Claude Code.