recursive-planning-spec

Expand feature intents into complete PlanGraphs with deterministic delta outputs.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/doctorduke/claude-config --skill recursive-planning-spec
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: recursive-planning-spec
Source: https://github.com/doctorduke/claude-config/tree/main/skills/recursive-planning-spec
Command: npx skills add https://github.com/doctorduke/claude-config --skill recursive-planning-spec

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill iteratively expands planning graphs from intent to terminal InteractionSpecs, ensuring completeness through nonterminal expansion, verification of proofs, and deterministic delta outputs.

Core Features & Use Cases

  • Nonterminal expansion to ensure leaf-level granularity
  • Deterministic deltas, proofs, and task ordering
  • UI projection for screens and components, plus supporting specs

Quick Start

Start a planning pass from a target feature, observe a delta shard, verify proofs, and iterate until completion.

Frequently Asked Questions about recursive-planning-spec

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

FAQPage Schema
How do I recursively decompose a feature into complete interaction specifications?

Recursive planning decomposes features from intent through nonterminal expansion into terminal InteractionSpecs. Start with a target feature, iteratively expand plan nodes, verify completion proofs (P1–P13), and generate delta-only artifacts until all leaves reach interaction-spec granularity.

What is nonterminal expansion in planning graphs?

Nonterminal expansion breaks down abstract planning nodes into more concrete child nodes until reaching terminal InteractionSpecs. It ensures leaf-level granularity, enforces OpenQuestion gating, and maintains deterministic task order with proof validation at each step.

How do I ensure my plan graph is complete and valid?

Completion-proof validation (P1–P13) verifies each node satisfies functional and technical requirements. Run proofs after each expansion pass, confirm delta outputs are deterministic, and iterate until all leaves are terminal InteractionSpecs with no unresolved OpenQuestions.

Can I use recursive planning to project UI implications across screens and navigation?

Yes. Recursive planning generates UI projections for screens, components, and navigation flows alongside InteractionSpecs. Delta outputs capture screen contracts and UI implications, enabling UI-centric software design within the same planning workflow.

What output formats does recursive planning produce?

Recursive planning outputs delta-format artifacts—only new or changed nodes—rather than full graph dumps. Outputs include deterministic task ordering, completion proofs, InteractionSpecs, and UI-projection metadata for screens and navigation contracts.

Does recursive planning work with MCP-based workflows?

Yes. Recursive planning integrates with MCP (Model Context Protocol) to drive nonterminal expansion and proof validation. It applies across AI planning and UI-centric design, enforcing delta outputs and guardrails against guessing at each decomposition step.