slate-plan

Reviews Slate v2 architecture plans against React 19.2 performance and regression gates.

16.5k|997|Updated Dec 2, 2019
One-click install
npx skills add https://github.com/udecode/plate --skill slate-plan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: slate-plan
Source: https://github.com/udecode/plate/tree/main/.agents/skills/slate-plan
Command: npx skills add https://github.com/udecode/plate --skill slate-plan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Repeated architecture review prompts for Slate v2 tend to produce endless incremental suggestions without a clear completion signal. This Skill turns those reviews into a scored, pass-scheduled plan with explicit closure gates so a plan only completes when every required pass, verification, and handoff gate is done.

Core Features & Use Cases

  • Scored plan review: Scores Slate v2 architecture and API plans across weighted dimensions including React 19.2 runtime performance, Slate-close unopinionated DX, migration-backbone shape, regression-proof testing, research evidence, and shadcn-style composability.
  • Two-phase lane workflow: Planning mode hardens the plan through scheduled passes; execution mode starts only after explicit user acceptance and runs against the live .tmp/slate-v2 workspace with recorded verification commands.
  • Issue and evidence ledgers: Tracks fixed and related GitHub issues, fork issue dossiers, PR reference sync, maintainer objection ledgers, and ecosystem strategy synthesis for Lexical, ProseMirror, and Tiptap evidence.
  • Use Case: When reviewing whether a proposed Slate v2 plugin API is the right shape for Plate and slate-yjs migration, invoke this Skill to produce a scored plan with cited evidence, decision briefs, and a closure checklist instead of another round of open-ended suggestions.

Quick Start

Ask the assistant to run slate-plan with a Slate v2 architecture review prompt, for example: review the proposed editor extension API plan and score it against the closure gates.

Frequently Asked Questions about slate-plan

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review a Slate v2 architecture plan with a confidence score?

Invoke slate-plan with your architecture review prompt. It creates a plan under docs/plans, runs scheduled review passes, and scores six weighted dimensions from 0.00 to 1.00. The plan closes only when the total score reaches 0.92 with no dimension below 0.85 and all closure gates pass.

What is the difference between planning mode and execution mode in slate-plan?

Planning mode edits only planning, research, and ledger artifacts and never touches Slate v2 implementation code. Execution mode starts only after you explicitly accept a ready plan and invoke the skill again, allowing edits to the .tmp/slate-v2 workspace with required verification commands.

Does slate-plan verify changes against the actual Slate v2 codebase?

Yes. Any claim about Slate v2 source, runtime, browser behavior, or issue fixes must be verified with commands run in the live .tmp/slate-v2 workspace, with the exact command, directory, and result recorded in the plan. Tests run elsewhere do not count as Slate v2 verification.

When should I not use slate-plan?

Do not use it for narrow bug fixes, browser reproductions, or normal diff code reviews. It is designed for architecture, public API, and migration planning lanes where repeated review keeps producing suggestions without a completion signal.

How does slate-plan handle Plate and slate-yjs migration concerns?

It requires substrate-level migration-backbone answers such as extension namespaces, schema policy, deterministic operations, and commit metadata. It does not require compatibility with current Plate public APIs or slate-yjs adapters, since raw Slate only needs to provide the architecture backbone.