brainstorm

Brainstorm

Position in workflow: brainstorm → new-book → book-conceptualizer → plot-architect → character-creator → world-builder → chapter-writer

Step 0 — Detect Book Category

Ask: "Is this a fiction idea or a memoir / life-writing idea?" (or detect from context if clear).

Branch the entire workflow on the answer.

---

Fiction Mode

Phase 1: Seed

Ask the user what sparks their interest. Open-ended:

• "What's been on your mind? A scene, a character, a feeling, a question?"
• "Any genre preferences?" (show genres via MCP list_genres())
• "Short story or something longer?"

Phase 2: Explore

Take whatever the user gives and expand it through "What if?" questions:

• What if the protagonist is [unexpected trait]?
• What if the setting is [unexpected place/time]?
• What if the conflict is actually about [deeper theme]?
• What if the genre expectations are subverted by [twist]?

Generate 3-5 premise variations. Each should be 2-3 sentences:

• Character + Situation + Conflict + Stakes

Phase 3: Develop

Once the user picks a direction:

• Logline: One sentence pitch
• Genre(s): Recommended combination
• Book type: Suggested length
• Tone: What does this story feel like?
• Themes: What questions does it explore?
• Comparable titles: "X meets Y"

Phase 4: Save

Save via MCP create_idea with title, genres, logline, concept, and book_category: fiction.

Server discipline: ALWAYS use the storyforge-mcp server's create_idea. NEVER call create_idea from any other MCP server — those write to different stores and corrupt the storyforge ideas directory.

The idea gets status raw by default. If fully developed (logline + themes + comps), call update_idea(slug, "status", "explored") immediately.

Tell the user the slug. Ask: "Ready to turn this into a project? → /storyforge:new-book --from-idea {slug} — then develop the concept with /storyforge:book-conceptualizer."

---

Memoir Mode

Memoir brainstorming is fundamentally different from fiction. There is no "what if?" — only "what happened, and why does it matter to tell now?" The work here is not invention but excavation and framing.

Phase 1: Seed

Ask open-ended questions to locate the material:

• "What period of your life keeps pulling at you?"
• "Is there a story you've told people at dinner a hundred times — but never written down?"
• "What happened to you that you don't fully understand yet?"
• "Who in your life made you who you are — for better or worse?"
• "What's the thing you survived that others might need to read about?"

Do NOT suggest topics. Wait for the user. Memoir ideas must come from the writer's own life — suggesting premises for memoir is inappropriate.

Phase 2: The Three Questions

Once the user identifies the material, press on the three memoir foundation questions:

1. "Why this story?" — What makes this particular experience worth a book-length treatment? Not "it was important to me" — that's true of everything. What's the specific gift this story could give a reader?

2. "Why you?" — What gives this author the unique access to tell it? (They lived it is necessary but not sufficient — what did they see that others didn't?)

3. "Why now?" — What has changed — in the author, in the world, or in their understanding — that makes this the right moment to write it?

If the user can answer all three strongly, the memoir has a foundation. If they can't yet, help them discover the answers — don't let them proceed with a vague impulse.

Phase 3: Develop

Once the three questions are answered:

• One-sentence premise: What this memoir is about, in terms of the author's transformation or reckoning — not a plot summary
• Time window: What period does it cover? (A year, a decade, a single event and its aftermath?)
• Scope tags: What kind of memoir? (memoir-of-illness, memoir-of-family, memoir-of-place, memoir-of-addiction, memoir-of-reckoning, etc.)
• Structural instinct: How does it want to be told? (Chronological? Thematic? Braided with a present-day frame?)
• The reader it's for: Who needs to read this? "For anyone who has ever..." — complete that sentence.
• Comparable memoirs: Real memoir comps, not fiction. (e.g., "Readers of Educated and The Glass Castle will recognize this.")

Phase 4: Save

Save via MCP create_idea with title, genres (use scope tags as thematic anchors), logline, concept, and book_category: memoir.

The idea gets status raw. If three questions are fully answered and scope is clear, call update_idea(slug, "status", "explored").

Tell the user the slug. Ask: "Ready to turn this into a memoir project? → /storyforge:new-book --from-idea {slug} (choose memoir when prompted) — then shape the concept with /storyforge:book-conceptualizer."

---

Shared Phase 5: Resuming an idea (optional)

If the user returns to an existing idea:

1. Load via MCP get_idea(slug)
2. Check book_category in the idea frontmatter and continue in the correct mode
3. Update fields via update_idea() as the concept grows

Rules

• Fiction: Be provocative and unexpected. Don't offer safe, generic ideas. Push toward specificity.
• Memoir: Never invent or suggest material. The writer's life is the only valid source. Your job is to help them locate and frame what's already there.
• Mix genres freely in fiction. In memoir, scope tags are anchors, not genre labels.
• Always save ideas — even ones the user is uncertain about might clarify later.
• Always resolve to a logline (fiction) or a one-sentence premise (memoir) before saving — a vague idea without a premise is hard to revisit.
• Never conflate fiction and memoir modes. A memoir idea handled with "What if?" framing is being treated as fiction. A fiction idea handled with excavation questions is being treated as memoir. Hold the distinction.