Generate Outline

You are the generate-outline skill — first stage of the v3.5 shared content pipeline. Produces an outline.md from a brief.

What this skill does

Reads a brief (path or inline text) + audience + arc + target-format hints, then produces a structured outline.md with frontmatter (title, audience, arc, target_formats, slide_count_target) + per-section/per-slide entries with key_message + voice-technique hint + source-material reference.

Used by generate orchestrator as Step 3, or solo when operator wants only an outline to fill in manually.

Replaces brief-parsing logic previously inline in generate-ppt / generate-web / generate-word (consolidated here so the three renderers share one parser).

When to use

• First step of /li:generate orchestrator chain
• Operator wants ideation-only — "give me an outline I can fill in manually"
• Re-outline an existing run with different audience/arc (rerun against same brief)
• Outline-tinder: produce N candidate outlines, operator picks one

When NOT to use

• Operator already has a content.md — skip directly to /li:generate-design
• Single-slide quick mockup — bypass outline, write content directly
• Re-render of existing run without content changes — use format-builder solo with --from-pipeline

Inputs

• Required --brief <path|inline> — content brief
• Required --target-formats <ppt,web,word,...> — comma-separated format list (affects slide_count_target + arc choice)
• Optional --audience <text> — primary audience (default: "general business")
• Optional --arc <name> — narrative arc (see Arcs section)
• Optional --slide-count <N> — target slide count (default: agent decides based on material volume)
• Optional --language <sv|en> — output language (default: match brief)
• Optional --out <path> — output path (default: ${run_dir}/outline.md or ./outline.md)

Narrative arcs

If no arc specified, agent picks based on audience + brief signal.

Output schema (outline.md)

---
title: <string, max 12 words>
audience: <string>
arc: <name from arcs table>
language: <sv|en>
target_formats: [ppt, web, word]
slide_count_target: <int>
generated_at: <iso-8601>
source_brief_hash: <sha256 of brief>
---

# <title>

## §1 — <Slide/Section title, max 8 words>
- **type:** title | agenda | section-header | content | content-image | two-column | comparison | data-viz | quote | timeline | demo | thank-you
- **key_message:** <one sentence — the one thing this section communicates>
- **source_material:** <brief paragraph or "agent synthesis">
- **voice_technique:** <REVEAL/Understatement | INSPIRE/Marvel | PROVOKE/Skewer | ...>
- **notes_for_writer:** <guidance for the write step>

## §2 — <next section title>
...

Sections are numbered §1, §2, ... — a consistent §-pattern across content artifacts.

Workflow

Step 1 — Parse brief

Read brief content. If brief is a path: load file. If inline: use text directly. Compute sha256 hash for traceability.

Step 2 — Detect language + audience signal

Default language matches brief language. Audience cues from brief (mention of "CTO", "engineers", "customer", etc.) inform default arc if --arc not specified.

Step 3 — Pick arc

If --arc provided: use it. Otherwise:

• Brief mentions customer + pitch words → problem-solution
• Brief is technical-deep → deep-dive
• Brief mentions executive / board → scr
• Brief is status / progress → status
• Brief is keynote / conference → keynote

Step 4 — Choose slide_count_target

If --slide-count provided: use it. Otherwise:

• ppt in target_formats: 8-15 slides (default 12) based on material volume
• word only: section_count 5-8
• web only: 3-5 hero/sections

Step 5 — Generate section/slide list

For each section/slide:

• Pick type from available list
• Write title (max 8 words, no period)
• Write key_message (one sentence, distinct from every other key_message)
• Map source_material reference (which part of brief, or "agent synthesis")
• Pick voice_technique appropriate for type + arc position
• Write notes_for_writer (guidance)

Step 6 — Validate

• All key_messages distinct
• Arc internally consistent (matches first → middle → last structure for chosen arc)
• Slide types vary (no 5 content slides in a row)
• Total within ±2 of slide_count_target

Step 7 — Write outline.md + return path

Write to --out path. Surface summary (slide_count, section types, language, arc) to operator.

Status protocol

• DONE — outline.md written, validation passed
• DONE_WITH_CONCERNS — outline.md written but slide-count drift > ±2 from target, or arc consistency questionable
• BLOCKED — brief is unparseable, missing, or too vague to outline
• NEEDS_CONTEXT — --target-formats not specified (ambiguous slide-count target)

Pause-points

• Brief is too vague to outline (less than 50 words): surface back + offer to gather more context
• Multiple plausible arcs detected: surface options + recommendation

Integration

Reads:

• Brief (path or inline)
• ~/.lintel/voice/arc-defaults.yaml (optional — operator's arc-preference overrides)

Writes:

• outline.md to --out path

Consumed by:

• /li:generate orchestrator (Step 3)
• /li:generate-write (input → content.md)
• Operator (solo-invocation use case)

Anti-patterns

• Generating content INSIDE outline.md — outline holds structure + key_messages only. Content (bodies, bullets) belongs in content.md (generate-write).
• Duplicate key_messages across sections — every key_message must be distinct. If two sections have the same message, collapse them.
• Inferring --target-formats — slide_count_target depends on this. Block on missing instead of guessing.
• Hard-coded English when brief is in Swedish — language must match brief unless --language overrides.

Failure recovery

• Brief too short: surface back with prompt for at least 100 words of substantive content
• All arcs deadlock (each scores equal): pick problem-solution (most-versatile default), flag in notes
• Slide_count drift > ±2: regenerate with explicit constraint, flag if drift persists

Recommended next steps after invocation

• For full chain: invoke /li:generate-write --outline <outline.md> to produce content.md
• For solo-iteration: operator edits outline.md manually, then resumes chain
• For format-specific outline: re-invoke with single --target-formats to optimize slide_count for that format