planwright
v0.34.0Autopilot for spec-driven development. The human signs off and merges; planwright flies the spec. Scales to a multi-spec fleet: an agentic software factory on Claude Code primitives.
By Diego RomeroLicense: MIT3 GitHub starsUpdated 9 hours ago
Directory evidence
- Runtimes
- Claude Code
- Parsed components
- 11 skill or MCP entries
- Source updated
- Aug 31, 2026
- Manifest status
- Canonical path parsed
The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →
Install planwright for Claude Code
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install planwright@agent-plugin-marketplacePaste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.
The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.
Get the source manually
git clone https://github.com/inkatze/planwrightClone the source repository, then follow its setup instructions to add the plugin to a compatible client. The repository root is the plugin root.
Plugin files
├── .claude-plugin/plugin.json├── skills/builder/SKILL.md├── skills/drain/SKILL.md├── skills/execute-task/SKILL.md├── skills/offload/SKILL.md├── skills/orchestrate/SKILL.md├── skills/polish/SKILL.md├── skills/resume/SKILL.md├── skills/self-review/SKILL.md├── skills/spec-draft/SKILL.md├── skills/spec-kickoff/SKILL.md└── skills/spec-walkthrough/SKILL.md
Included Skills11
Detect a project's stack and recommend or apply the universal mechanical quality guards from planwright's core catalog (formatter, linters, type-checker, test runner, secret scan, commit hooks, CI gate), plus the growable breadth dimensions. Escalates stake-bearing decisions (auth, data modeling, security posture, integration surface) into the deferral mechanism instead of auto-defaulting them. Plugs into /spec-draft's design phase and /execute-task's guard-application step.
Run the on-demand drain pass over every spec bundle's Gate deferral entries: evaluate structured GATE(when:) conditions, surface date and free-text gates, report malformed ones, inventory each live bundle's [manual] test-spec entries, and surface the observations log's unmined state. Read-only; nothing is auto-resolved or auto-dropped.
Implement one task (or a cohesion bundle) from a signed-off spec (Ready or Active): recompute the execution freshness gate, write the verifying test first, implement to green, run the project's full CI with adaptive retry, converge via the configured review_sequence (default /polish --nested), then open a draft PR referencing the brief, tasks, REQs, and tests. The execution workhorse of the planwright pipeline. Assumes the worktree already exists; never creates worktrees, never merges, never marks a PR ready.
Offload a free-form petition of work to the right execution rung: apply the work-placement axioms (tower-frugality, smallest-sufficient-rung) to pick the smallest sufficient backend from what the host advertises, ask when the petition under-determines the choice, dispatch through the backend seam, and report the worker's handle with an observe/attach hint. The sole home of adaptive backend selection.
Advance one planwright spec by one step: pick the next ready unit critical-path-first (or a cohesion bundle), run the freshness gate, record the dispatch under the per-spec lock, and dispatch /execute-task via the backend. A stateless, disposable control tower; a reconcile sweep rebuilds from disk. Never merges, marks a PR ready, or auto-chains into /spec-kickoff. --bookkeeping runs the drain + PR reconcile; --watch loops the step.
Autonomous act-then-review convergence loop: iterate /self-review passes, draining every action disposition (Auto-applicable and Agent-resolvable applied, Needs-sign-off applied on the branch) until only irreducible Needs-human-judgment forks remain, then hand off the full audit record. Local-only: never pushes, never creates a PR. Pass --nested when invoked from a parent skill (such as /execute-task) that owns the handoff.
Load context for a fresh session in a worktree with in-flight pair-flow work. Reads the kickoff brief (the contract), tasks.md state, recent git log, the open PR state, and an optional handover brief, then surfaces uncommitted changes and asks before proceeding. Read-only over your in-flight work: no stash, clean, push, merge, or commit of it. The skill's only write is the standard self-healing drift-log note.
Gate-wired review pass of the current feature branch: Discovery Rigor fan-out, multi-pass validation, act-then-review dispositions per the finding categorization, full audit record (lens-coverage table, four bucket tables, declined log, pending-sign-off checklist). Standalone runs push and open or update a draft PR; pass --nested to stay local and return the audit record to the invoking skill.
Interactively elicit a four-file spec bundle (requirements.md, design.md, tasks.md, test-spec.md) at Status Draft on the spec's own branch, mining seed sources (pending notes, the observations log, transcripts) and citing them. Runs fold-detection against existing specs on every invocation and surfaces an extend recommendation instead of spinning a duplicate bundle. Commits the completed bundle (commit_on_draft opt-out); never pushes, never flips a spec to Ready.
Walk a spec bundle section by section to mutual understanding, producing the signed-off kickoff brief downstream skills execute from. On sign-off: runs the Discovery-Rigor lens pass, flips Draft to Ready, records the sign-off record (anchor last), commits, pushes, opens a draft PR, then on clean completion marks the spec PR ready (mark_spec_pr_ready_on_kickoff opt-out). Also runs delta re-walkthroughs and amendments; halts on genuine spec inconsistency rather than papering over it.
Render a spec bundle (or a chosen slice) into a plain-language, didactic comprehension artifact a human reads and judges for themselves: an unaided cold read before kickoff, re-orientation mid-execution, or onboarding to a finished or abandoned spec. Standalone and strictly read-only: it renders any status, never edits, signs off, mutates the pipeline, or delivers a verdict of its own. The independent complement to /spec-kickoff's guided dialogue.
Plugin manifests1
{
"$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
"name": "planwright",
"displayName": "Planwright",
"description": "Autopilot for spec-driven development. The human signs off and merges; planwright flies the spec. Scales to a multi-spec fleet: an agentic software factory on Claude Code primitives.",
"version": "0.34.0",
"author": {
"name": "Diego Romero",
"email": "[email protected]"
},
"license": "MIT",
"homepage": "https://github.com/inkatze/planwright",
"repository": "https://github.com/inkatze/planwright",
"keywords": [
"spec-driven-development",
"software-factory",
"agentic-coding",
"agent-orchestration",
"harness-engineering",
"loop-engineering",
"human-in-the-loop",
"guardrails",
"agentic",
"ai-agents",
"claude-code"
],
"defaultEnabled": true
}For maintainers
If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.
[planwright on Agent Plugins Marketplace](https://pluginsmp.com/plugins/planwright)