ideatelisted

# ideate — Stage 1 of the Spark lifecycle `Ideate → Plan → Codify → Validate → Ship` Ideate is where a fuzzy idea becomes a problem worth solving, written down. No code, no file layout, no tickets yet. The output is a short problem statement the `plan` skill can decompose. ## Do this 1. **Restate the idea in one sentence.** If you can't, the idea isn't ready — keep asking until you can. 2. **Survey prior art and existing assets.** Before framing the problem as new, check for a predecessor repo, a prototype, captured data, or an abandoned branch. Record what exists, what's reusable, and how it relates to this effort. For a rewrite, this is the highest-leverage question you can ask. 3. **Pressure-test it.** Invoke the **`grill-me`** skill (Claude-native) to interview the user down the decision tree. Resolve the load-bearing unknowns before writing anything. 4. **Write the problem statement.** Keep it to one screen: - **Problem** — what hurts today, for whom. - **Outcome** — what "solved" looks like, in observable terms. - **Success criteria** — 2–5 checks you could actually verify. - **Prior art & reusable assets** — what already exists, what carries over, what's deliberately left behind. - **Constraints** — stack, deadline, must-use / must-avoid. - **Non-goals** — what this explicitly will not do. 5. **Confirm.** Read it back. Get a yes before handing off to `plan`. 6. **Persist it.** Write the confirmed statement to `docs/problem-s