draft-post

name	draft-post
description	Draft the actual copy for a specific LinkedIn post (1 of 5 in the series) — produces 3 style variants (technical, business, founder), 5 opening hooks, and hashtags. Use this AFTER series-plan when the user wants to write the real text of post N. Requires the series plan from series-plan.
argument-hint	<post-number 1-5>

draft-post

For a specified post number (1–5) in the LinkedIn series, generate 3 distinct style variants (technical, business, founder), 5 opening hooks, and a hashtag set.

When to use

Run after /linkedin-content:series-plan. Run once per post in the series — you typically run this 5 times for a 5-post campaign.

User input

$ARGUMENTS

$ARGUMENTS must be a single integer 1–5 (the post number).

Argument parsing

• No argument: Output "ERROR: Post number required. Usage: /linkedin-content:draft-post N where N is 1–5." and STOP.
• Not 1–5: Output "ERROR: Invalid post number '$ARGUMENTS'. Must be 1–5." and STOP.

Prerequisites

Verify these files exist:

1. content/series-plan.md
2. content/claims.md

If series-plan.md is missing, output and STOP:

ERROR: Missing prerequisite.
- Series plan not found at content/series-plan.md
- Run /linkedin-content:series-plan first to generate it.

If claims.md is missing, output and STOP:

ERROR: Missing prerequisite.
- Claims not found at content/claims.md
- Run /linkedin-content:proof-pack first to generate it.

Process

Step 1: Load context

1. Read content/series-plan.md — extract Topic N's structure (title, core message, CTA angle, contrarian hook, supporting claims, repo artifacts).
2. Read content/claims.md — locate the claims that ground this topic. Require at least 2 supporting claims; if fewer, emit a WARNING but proceed.
3. If content/.config.md exists, read it for forbidden-phrase or tone overrides.

Step 2: Generate 3 post variants

Each variant MUST:

• Be 150–300 words (count carefully)
• Include at least 1 specific repo fact (real file, class, or design decision)
• Include at least 1 business outcome (concrete benefit in business terms)
• Open with a strong first line that creates curiosity and clarity
• Close with a soft CTA aligned with the topic's CTA angle
• Contain NO forbidden phrases: "AI is changing everything", "the future of", "game-changer", "revolutionary", "next-generation" (plus any from .config.md)
• Reference at least 1 claim from claims.md

Variant styles

1. Technical (draft-technical.md): Architecture-focused. Leads with how the system works. Precise technical language. Appeals to developers and architects.
2. Business (draft-business.md): Problem-solution, outcome-focused. Leads with the business problem this solves. Business language with just enough technical detail to be credible. Appeals to engineering managers and decision-makers.
3. Founder (draft-founder.md): Story-driven, community-focused. Leads with the journey or motivation behind the project. Authentic, personal voice. Appeals to open-source enthusiasts and the developer community.

Step 3: Generate 5 opening hooks

Each hook is 2–3 lines long, technically accurate, no superlatives, distinct framing:

1. Pain-first — start with a frustration the audience recognizes
2. Misconception-first — challenge a belief the audience holds
3. Architecture-first — lead with a technical decision that surprises
4. Business-first — lead with a concrete business outcome
5. Curiosity-first — pose a question that demands an answer

Step 4: Generate hashtag set

Total 5–7 hashtags (LinkedIn sweet spot). Mix:

• Domain hashtags (2–3): grounded in the project's language/platform (e.g., #TypeScript, #Salesforce, #Python, #OpenSource)
• Topic-specific (2–4): grounded in the post's specific angle (e.g., #APIDesign, #OAuth, #DeveloperExperience, #AIAgents)

Format as a single copy-paste line.

Step 5: Write output

Write 5 files for post N. Replace N with the actual number.

File 1: content/posts/post-N/draft-technical.md

# Post N — Technical: [Topic Title]

**Topic**: [N] — [Title]
**Variant**: Technical
**Word count**: [N]
**Claims referenced**: [C-XXX, C-YYY]

---

[Full post text]

File 2: content/posts/post-N/draft-business.md

# Post N — Business: [Topic Title]

**Topic**: [N] — [Title]
**Variant**: Business
**Word count**: [N]
**Claims referenced**: [C-XXX, C-YYY]

---

[Full post text]

File 3: content/posts/post-N/draft-founder.md

# Post N — Founder: [Topic Title]

**Topic**: [N] — [Title]
**Variant**: Founder
**Word count**: [N]
**Claims referenced**: [C-XXX, C-YYY]

---

[Full post text]

File 4: content/posts/post-N/hooks.md

# Hooks: Post N — [Topic Title]

**Topic**: [N] — [Title]

## 1. Pain-First

[2–3 lines]

## 2. Misconception-First

[2–3 lines]

## 3. Architecture-First

[2–3 lines]

## 4. Business-First

[2–3 lines]

## 5. Curiosity-First

[2–3 lines]

File 5: content/posts/post-N/hashtags.md

# Hashtags: Post N — [Topic Title]

**Topic**: [N] — [Title]

## Copy-paste block

[single line of 5–7 hashtags]

## Breakdown

- **Domain**: [hashtags + rationale]
- **Topic-specific**: [hashtags + rationale]

Tip: paste with formatting

To paste a draft into LinkedIn with bold/italic preserved:

pandoc content/posts/post-N/draft-business.md -t html | "${CLAUDE_PLUGIN_ROOT}/scripts/pbcopy-html.swift"

Error handling

Condition	Response
Missing prerequisites	HARD BLOCK — see Prerequisites
Post number not 1–5	ERROR with usage instructions
No $ARGUMENTS provided	ERROR with usage instructions
Topic has fewer than 2 supporting claims	WARNING: "Topic [N] has only [X] supporting claims. Consider re-running /linkedin-content:proof-pack to expand coverage." Proceed with available claims.