By mattmccray
A coarse, resumable program layer that sequences Phases above your change-management workflow (OpenSpec by default). The epic layer your change tool doesn't have.
Mark a deliverable/phase done — flip the checkbox, stamp status + date, reconcile
Finalize a completed plan — move it to plans/archive/, optionally chain into a new plan
Discuss a program, then scaffold a fresh PLAN.md (guards an incomplete active plan first)
Identify the next phase/deliverable and hand off to the change engine to carry it
Show where the plan stands — current phase, phase states, carrying-change states, and what's next
Runs pre-commands
Contains inline bash commands via ! syntax
Bash prerequisite issue
Uses bash pre-commands but Bash not in allowed tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A coarse, resumable program layer for Claude Code — the epic/roadmap layer that sits above a change-management workflow (OpenSpec by default).
Tools like OpenSpec take a single unit of work from proposal → implementation → archive, but have no concept of a program: a phased roadmap spanning many changes. Plan fills that gap with one human-readable file — PLAN.md — that sequences Phases, each carried by one or more Changes.
PLAN.md ............ the roadmap: Phases, sequenced, each carried by Changes
│ carried by
▼
openspec/changes/ .. Changes (propose → apply → archive) — the work engine
Plan is deliberately coarse: it's a spine, not a task tracker. The granular tasks
live only in the change that carries a phase — the plan points at them, never
restates them, so the two never drift. The document is the state: everything
needed to resume lives in PLAN.md itself.
Plan is a single-repo Claude Code plugin. The repo is its own marketplace:
/plugin marketplace add mattmccray/plan # or the full GitHub URL
/plugin install plan@plan-tools
To develop against the live repo instead, see DEV.md.
/plan:new "Minerva v3" # discuss the program, scaffold PLAN.md
/plan:next # find the next phase, hand off to opsx:propose
/plan:advance # land a phase: flip the checkbox, stamp status
/plan:status # where are we? (plan ✕ real change state)
/plan:archive # finalize a complete plan → plans/archive/
You can also just talk to it — "let's start a new plan", "where are we?",
"what's next?" — and the plan-workflow skill routes to the right procedure.
| Command | Description |
|---|---|
/plan:new | Discuss a program, then scaffold a fresh PLAN.md (guards an incomplete active plan first) |
/plan:status | Show current phase, phase states, carrying-change states, and what's next |
/plan:next | Identify the next phase/deliverable and hand off to the change engine |
/plan:advance | Mark a deliverable/phase done — flip the checkbox, stamp status + date |
/plan:archive | Finalize a complete plan → move to plans/archive/, optionally start the next |
/plan:validate | Drift check: plan checkboxes ↔ real change state |
PLAN.md); one active plan per program.A Plan sequences Phases; each Phase is carried by one or more Changes; Invariants hold across all Phases; Divergences are recorded, never silent.
PLAN.md at the repo root.plans/archive/YYYY-MM-DD-<program>.md, date-prefixed so
they pair with OpenSpec's openspec/changes/archive/. Completed plans are
kept, not deleted — they're the historical record of why a program was
built the way it was.Plan does not hard-code OpenSpec: the plan declares its change engine in its header, and the workflow drives whatever is declared (defaulting to OpenSpec). It is one tool in a constellation of single-purpose, AI-first workflow tools — adopt it alongside a change engine and (optionally) an issue inbox, picking only what a project needs. No central store, no dispatcher; each tool owns its own file and its own slash namespace.
See DESIGN.md for the full ratified model.
MIT
npx claudepluginhub mattmccray/plan --plugin planPersistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows
Unity Development Toolkit - Expert agents for scripting/refactoring/optimization, script templates, and Agent Skills for Unity C# development