ideate

An Agent Skill — for Claude and OpenAI Codex — for turning a fuzzy idea (or an existing project you want to improve) into a locked concept + phased roadmap — captured in one living CONCEPT_BRIEF.md. It's the upstream partner to the prompt-pack skill:

ideate → CONCEPT_BRIEF.md → prompt-pack → execute

It encodes a method reverse-engineered from real idea→ship journeys — including the mistakes those journeys made, so it improves on them instead of repeating them.

What it does

Runs a disciplined funnel as a blunt, honest co-founder:

1. Frame — absorb the brain-dump, set the "what makes this worth doing" rubric, and force the things people always skip: a real success metric, a kill criterion, and — when feel is the wedge — how load-bearing the design is (a substantive design direction carried into the brief).
2. Explore — greenfield: 5–7 comparable options ranked to your criteria, then mine your own lived experience for the wedge. Refinement: a four-lens assessment (what is it / viability / quality / honest take) and narrow strategic forks.
3. Pressure-test (the gate) — two-sided honest assessment, competitor buckets, viability economics fit to the concept's type (market + unit economics if it's commercial; reach + cost-to-sustain if it's free/OSS/internal — cited fact vs estimate), feasibility tiers, a red-team pass, and verification against real artifacts. For high stakes it delegates to the deep-dive skill. Ends in a verdict and a go / iterate / kill decision — and before any kill it steelmans the idea's best case and frames the call as park-with-a-revisit-trigger, so it won't talk you out of a good idea on a confident hunch.
4. Converge — lock decisions one at a time (with a decision log), scope IN/OUT with deferrals named, lock the spine and defer the rest.
5. Handoff — name it late, then offer to hand the brief to prompt-pack.

Its central mechanism is one living CONCEPT_BRIEF.md, edited in place, never regenerated — the fix for the #1 failure in unstructured ideation: the AI re-deriving the same assessment every session while decisions never settle.

Two modes: greenfield (fuzzy idea → concept) and refinement (evaluate/improve an existing thing, treating it as the "golden master" with an enforceable parity contract).

Quickstart — try this first

You type: "I have an idea for — help me decide if it's worth building."

You get back: a living docs/CONCEPT_BRIEF.md — a locked concept, a forced success metric and kill criterion, and a phased roadmap. ideate will pressure-test the idea (and push back) before it writes a line of roadmap.

Install

This is an open Agent Skill — the same ideate/ folder works in Claude and OpenAI Codex. Pick your tool:

You use…	Install it by…
Claude Code — terminal, the Code tab of the Claude desktop app, claude.ai/code, or a VS Code / JetBrains IDE	dropping ideate/ into ~/.claude/skills/ (all projects) or .claude/skills/ (one project)
OpenAI Codex — CLI, app, or IDE	dropping ideate/ into ~/.agents/skills/ (all repos) or .agents/skills/ (one repo), then restarting Codex
Claude chat — the Chat tab of the desktop app, or claude.ai	uploading ideate.skill (the zip in this repo) under Customize → Skills
Any other agent	pointing it at ideate/SKILL.md — it's just instructions

The .skill file is just the ideate/ folder zipped, so one unzip drops it into either skills home:

# Claude Code — detected in-session (verify with /skills):
unzip ideate.skill -d ~/.claude/skills/

# OpenAI Codex — the current skills path is ~/.agents/skills/; restart Codex after:
mkdir -p ~/.agents/skills && unzip ideate.skill -d ~/.agents/skills/

Prefer a clone? git clone https://github.com/nelsonwerd/ideate-skill.git, then cp -r ideate-skill/ideate into whichever skills folder above.

_{Menu names and exact paths shift between versions — the Claude Skills and Codex Skills docs are the source of truth. Auto-activation by description is native to Claude; in Codex you invoke the skill explicitly. The methodology is identical across both.}

Runtime support

	Claude chat	Claude Code	OpenAI Codex	Other agents
ideate	Strong — concept work; brief kept inline when there's no file tree	Best	Strong — with a local workspace for the brief	Works — full method; keep the brief in a file or inline

Use it

• Manually: type /ideate and describe the idea (or point it at an existing project to evaluate).
• Automatically: Claude invokes it when you say things like "help me figure out what to build", "is this idea any good?", "should I rebuild X?", "where are we really at?", or "turn my idea into a plan."

Examples:

• "I have an idea for a scheduling tool for tutors — is it any good?"
• "Here's my repo. Deep-dive it: viability, quality, honest take — should I keep going?"
• "Help me figure out what to build; I want a subscription SaaS but don't have a lane yet."

It pairs directly with prompt-pack: when the concept is locked, ideate hands the CONCEPT_BRIEF.md to prompt-pack, which turns the roadmap into sequenced, self-contained build prompts.

What's in this repo

• ideate/ — the skill itself (SKILL.md + reference playbooks). This is what you install.
  • references/concept-brief-template.md — the canonical CONCEPT_BRIEF scaffold (the core output).
  • references/facilitation-guide.md — the honest-advisor stance + the pressure-test battery.
  • references/modes-guide.md — greenfield vs refinement, in detail.
  • references/handoff-guide.md — the exact contract for handing off to prompt-pack.
• ideate.skill — the same folder, zipped, for one-step install.