vision-to-plan

Elicit a system steward's vision into approved constitution, PRD, and actors governance documents.

Updated Apr 19, 2026
One-click install
npx skills add https://github.com/nickgogan/improvement-loop --skill vision-to-plan-nickgogan
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vision-to-plan
Source: https://github.com/nickgogan/improvement-loop/tree/main/.claude/skills/vision-to-plan
Command: npx skills add https://github.com/nickgogan/improvement-loop --skill vision-to-plan-nickgogan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Turning a system steward's rough vision into durable, agent-executable planning documents is slow and inconsistent. This Skill structures that process through Socratic elicitation and section-by-section human approval, producing governance artifacts a context-free agent can act on. ## Core Features & Use Cases - Governance-first artifact chain: Produces a confirmation summary, constitution, PRD, and actors document in sequence, each approved section by section before the next begins. - Elicitation templates: Ships constitution, PRD, and actors templates with embedded per-section interview instructions that are removed from final artifacts. - Milestone handoff: Maps every approved epic to PROGRESS.md milestone proposals with binary definitions-of-done, in the session-handoff vocabulary. - Use Case: A steward wants to re-plan a subsystem after a direction change. The Skill diffs the new intent against existing CHARTER and governance docs, proposes section-scoped amendments, and records superseded decisions rather than silently rewriting them. ## Quick Start Ask the agent to capture the vision for your system and turn it into a constitution, PRD, and actors document with section-by-section approval.

Frequently Asked Questions about vision-to-plan

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

FAQPage Schema
How do I turn a rough product vision into a PRD?▼

Invoke the skill with a request like "create a PRD" or "capture the vision". It interviews you section by section, drafts the constitution, PRD, and actors documents from templates, and writes each section to governance files only after your explicit approval.

What documents does vision-to-plan produce?▼

It produces three kernel documents in the governance directory: constitution.md (or a CHARTER amendment set), prd.md with epics and binary acceptance criteria, and actors.md defining agent boundaries and permissions. It also proposes PROGRESS.md milestone entries.

Can vision-to-plan run unattended or in batch mode?▼

No. An interactive per-section human-approval channel is a required capability; the skill is explicitly inert without one. Unattended runs are out of scope because the section gate is the core method.

When should I not use vision-to-plan?▼

Do not use it for end-of-session reconciliation (use session-handoff), skill authoring (use design-skill or meta-skill-author), research intake, or single small tasks with an obvious plan, which should be planned inline.

How does vision-to-plan handle existing planning documents?▼

Re-entry is the default posture: it diffs the steward's new intent against existing CHARTER, constitution, PRD, and actors files, proposes section-scoped edits, and records superseded decisions in append-only decision logs rather than silently rewriting.