OpenPlanr gives your coding agent 27 skills for specifying, planning, reviewing, designing, diagramming, shipping, and operating work from plans stored in your repository under .planr/. This package is the deterministic half: the planr command that stores and validates those files, renders diagrams and reports, installs the skills into each host, diagnoses the installation, and syncs with GitHub Issues and Linear. It never calls a model. Reasoning happens in the agent; the CLI keeps the record straight.

Install

npm install -g openplanr
planr --version

Also available as openplanr and opr. Alternatives: npx openplanr@latest setup, or the installers curl -fsSL https://openplanr.dev/install.sh | sh and irm https://openplanr.dev/install.ps1 | iex. Requires Node.js 20 or later.

Quick start

planr setup --runtime claude --scope user
# Codex:  planr setup --runtime codex --scope user --skill-mode unified-plugin
# Cursor: planr setup --runtime cursor --scope project

cd your-project
planr init
planr rules generate --target claude   # adds the skill map to CLAUDE.md (or codex → AGENTS.md)

Restart the coding agent, then ask for a specification: /planr:spec "…" in Claude Code, $planr:spec "…" in Codex, or mention the planr-spec rule in Cursor. Continue with plan to decompose it into stories and tasks and ship to implement one task. Plan and ship stay separate steps that you invoke.

planr setup --dry-run previews every file before anything is written; existing files are backed up byte for byte and planr runtime rollback restores them. See the setup guide.

What the CLI does

| Group | Commands | Purpose | | --- | --- | --- | | Planning files | spec, epic, feature, story, task, quick, backlog, sprint, template, checklist, update, search, graph | Create, validate, and query artifacts under .planr/ in three postures: spec-driven (SPEC-NNN-{slug}/), agile hierarchy, and quick tasks | | Hosts | setup, runtime, rules generate, doctor, upgrade | Install and update the skills for Claude Code, Codex, and Cursor; generate host guidance; diagnose drift; compare with the published set | | Delivery | status, sync, land, export, report, dashboard | Delivery status across specs and backlog, cross-reference repair, landing plans, stakeholder reports, and the loopback-only planning dashboard | | Design and diagrams | artifact, diagram | Review any HTML artifact with pins and decisions; render and verify diagrams offline across 39 grammars | | Operations and trackers | operate, github, linear, company | Run durable Operate cycles; two-way sync with GitHub Issues and Linear; publish selected artifacts to a company workspace |

Every command supports --yes for non-interactive use and most support --json. The CLI reference documents every command and flag.

Skills run in your agent

The skills (spec, plan, plan-review, sprint, ship, browser-qa, design, design-loop, design-review, diagram, artifact, land, release, doctor, investigate, status, sync, dashboard, openplanr, operate, and the seven Operate reviews) ship inside this package and are installed by planr setup. The skill catalog lists each skill's triggers, deferrals, and packaged references.

| Host | Install | Invoke | | --- | --- | --- | | Claude Code | planr setup --runtime claude --scope user | /planr:<skill>; ship dispatches nine role agents | | Codex | planr setup --runtime codex --scope user --skill-mode unified-plugin | $planr:<skill> | | Cursor | planr setup --runtime cursor --scope project | mention the planr-<skill> rule in Composer |

Documentation

• CLI reference
• Cross-runtime setup, migration, and rollback
• Troubleshooting
• Artifact review and private sharing
• Project documentation, getting started, and support

License

MIT. The OpenPlanr name and logo are covered by the trademark policy.