flow-next-chart

Resolves one decision at a time to map oversized unclear ideas into capture-ready briefings.

692|55|Updated Dec 26, 2025
One-click install
npx skills add https://github.com/gmickel/gmickel-claude-marketplace --skill flow-next-chart
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flow-next-chart
Source: https://github.com/gmickel/gmickel-claude-marketplace/tree/main/plugins/flow-next/codex/skills/flow-next-chart
Command: npx skills add https://github.com/gmickel/gmickel-claude-marketplace --skill flow-next-chart

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Large, ambiguous product ideas are too big to capture as a single spec, and jumping straight to implementation causes agents to drift, forget requirements, and ship unreviewable work. This Skill breaks one oversized idea into a decision map and resolves one decision per invocation until the effort can be handed off as a briefing.

Core Features & Use Cases

  • Decision-map discovery: Grounds a free-form idea against the repository, proposes an outcome, a visible frontier of decisions, and parked unknowns, then persists the chart via the bundled flowctl CLI.
  • Typed evidence routes: Resolves decisions through research, probe, eval, prototype, interview, or task routes, with attended decisions hard-gated so unattended drivers never fabricate human answers.
  • Briefing handoff: Clusters resolved decisions into one or more spec proposals and emits a briefing for the capture step, without ever writing specs itself.
  • Use Case: You have a vague multi-tenant migration idea. Chart it, let unattended research decisions resolve in parallel invocations, answer the attended interview decisions yourself, then receive a briefing ready for spec capture.

Quick Start

Ask the agent to chart out your oversized idea, for example by saying "chart out the multi-tenant migration" and then approve the proposed decision map.

Frequently Asked Questions about flow-next-chart

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

FAQPage Schema
How do I break down an idea too big for a single spec?

Invoke the chart skill with the free-form idea. It runs a bounded grounding snapshot, proposes an outcome plus the smallest visible frontier of decisions, and persists the chart through flowctl after your read-back approval, resolving nothing in that session.

How does decision-map discovery work before writing a spec?

Each work invocation claims exactly one decision from the live frontier, runs its typed evidence route such as research, probe, eval, prototype, or interview, resolves it with evidence, and recomputes the frontier. When no open decisions remain, a briefing is emitted for capture.

Can unattended agents resolve attended decisions like interviews?

No. Attendance is stored per decision and validated by flowctl. An unattended driver reaching an attended decision writes no answer, releases or leaves the claim visible, and terminates with a NEEDS_HUMAN verdict.

When should I not use decision charting for an idea?

Skip charting when the idea names a direction rather than a nameable destination, or when grounding shows no consequential unknowns. In those cases the skill creates nothing and recommends narrowing the prompt, running prospect, or going straight to capture.

What happens when a prototype reverses an earlier decision?

The new decision is resolved with a supersedes reference to the invalidated decision after a consent read-back. Superseded ledger lines stay struck-through rather than deleted, and dependent claims cascade according to the recorded graph.