create-plan

Orchestrates plan creation with per-chapter authoring and iterative validation loops.

Updated Mar 21, 2026
One-click install
npx skills add https://github.com/mikestopcontinues/agent-skills --skill create-plan-mikestopcontinues
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: create-plan
Source: https://github.com/mikestopcontinues/agent-skills/tree/main/claude/skills/create-plan
Command: npx skills add https://github.com/mikestopcontinues/agent-skills --skill create-plan-mikestopcontinues

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning an approved spike or feature idea into a validated, chaptered implementation plan requires coordinating scaffolding, multi-reviewer validation, feedback triage, revision, and decision locking — a process that is error-prone and easy to shortcut when done ad hoc. ## Core Features & Use Cases - Plan scaffolding: Creates a new plan via the yolo new plan CLI with auto-assigned pNN- numbering inside the correct project subtree. - Per-chapter validate-loop: Authors each chapter, then runs validate-doc, triage-feedback, revise-doc, process-feedback, and lock-decisions until convergence (minimum 2 iterations, hard cap at 3). - Overall README pass: Validates the plan README end-to-end after all chapters converge, with captain approval gating any chapter reopening. - Use Case: After a spike on an auth refactor is approved, invoke this skill to scaffold the plan, author chapters in dependency order, run the baseline reviewer set (accuracy, architecture, integration, dx, scope), and produce a convergence summary with locked decisions and pending clusters. ## Quick Start Ask the agent to create an implementation plan named auth-refactor from the approved spike, and it will scaffold the plan, author and validate each chapter to convergence, and report a convergence summary.

Frequently Asked Questions about create-plan

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

FAQPage Schema
How do I create an implementation plan from an approved spike?▼

Invoke the create-plan skill with a kebab-case plan name and the source spike path. It scaffolds the plan via the yolo CLI, authors chapters in dependency order, and runs the validate-loop per chapter until convergence.

What reviewer focus areas does plan validation use?▼

The baseline focus list is accuracy, architecture, integration, dx, and scope. The orchestrator adds contextual focuses such as security for auth or crypto chapters, performance for hot paths, and compatibility for public surfaces.

When should I use create-plan instead of create-spike?▼

Use create-plan when there is an implementation target requiring an ordered task breakdown, typically after a spike is approved. Use create-spike for pure research with no implementation target.

How many validation iterations does each plan chapter go through?▼

Each chapter runs a minimum of two iterations, even after a clean first pass, with a hard cap at three. Remaining clusters after iteration three are surfaced to the captain in the convergence summary rather than looping further.

What happens to unresolved feedback clusters after plan validation?▼

Every unresolved cluster appears in the mandatory convergence summary with its route, severity, and current state. Hiding pending clusters is explicitly forbidden, and decision-shaped clusters are routed to lock-decisions for captain review.