Create LinkedIn Content

Draft LinkedIn post variants that sound like a real human practitioner, not a LinkedIn thought-leader persona. Reads a voice guide the user has already generated (or prompts to create one), produces multiple framings of the same idea, and saves each variant as its own markdown file with frontmatter.

This is an agent-executed skill — the agent does the drafting and self-check inline. No Python script.

Mirrors create-x-content structurally; differences are marked ★.

Quick Start

/create-linkedin-content --brief "New open-source CLI that turns Figma files into React components. Called figma2react. Free, MIT licensed."

Or interactively:

/create-linkedin-content

Inputs

Voice Guide Resolution

Resolve in this order, stop at first hit:

1. --voice-guide <path> flag
2. ~/.goose-skills/config.json → voice_guides.linkedin
3. ~/.goose-skills/voice-guides/voice-linkedin.md (default path)
4. Fallback prompt — no guide found. Three options:

• (a) Paste a path to an existing guide
• (b) Run /generate-voice-guide --platforms linkedin now to create one (recommended)
• (c) Proceed with a neutral default (warn that variants will sound generic)

LinkedIn-Specific Defaults ★

Apply these unless the voice guide explicitly says otherwise:

• Length: 150–500 words per variant (LinkedIn rewards longer than X but shorter than a blog post)
• Bullets: arrow style (→) for workflow steps — this is LinkedIn-native
• Numbered steps: emoji numbers 1️⃣ 2️⃣ 3️⃣ allowed sparingly for installation/quickstart type posts
• "Why this matters" beat: every post should have one — LinkedIn audiences want the so what? more explicitly than X audiences
• Hashtags: 0–2 max. Prefer zero.
• CTAs: "Link in comments" pattern is LinkedIn-native; inline links are fine too. Avoid "Follow me for more."
• Openings: more scene-setting is ok (1–2 lines of context before the hook). Unlike X, you don't have to hook-first.

Workflow

Phase 1 — Load the voice guide

Same as create-x-content Phase 1 — load voice guide, extract banned phrases, hook patterns, format rules, dos/don'ts, examples.

Phase 2 — Parse the brief and decide variant count

Same 2–5 decision rubric. LinkedIn briefs often warrant fewer variants than X because LinkedIn posts are longer and the audience expects more narrative coherence — 3 strong framings usually beats 5.

Phase 3 — Generate variants with LinkedIn-appropriate framings

Framings tuned for LinkedIn (subset of the X framings, plus a few LinkedIn-native ones):

Phase 4 — Self-check against the voice guide

Same 4 checks as create-x-content:

1. Banned phrase check
2. Length check (150–500 for LinkedIn)
3. "Meat" check — specific tools/numbers/builds
4. Voice-match spot check

Plus LinkedIn-specific:

1. Arrow-bullet check — if the post uses lists, are they → style (or numbered emojis), matching the LinkedIn convention?
2. "Why this matters" check — does the post explain the stakes, or does it just state the facts? Add a "why this matters" beat if missing.
3. No corporate tone — does it sound like a press release? If yes, rewrite to sound like a human practitioner.

Phase 5 — Save outputs

File naming (the linkedin- prefix distinguishes from X variants when they live in the same folder):

linkedin-<letter>-<framing-slug>.md

Examples: linkedin-a-builder-story.md, linkedin-b-product-launch.md.

Frontmatter:

---
id: <topic-slug>-li-<letter>
platform: linkedin
format: standard | long
topic: <slug>
framing: <framing-slug>
status: draft
---

Phase 6 — Deliver

Print output directory, file list, framing summary, suggested next step.

Outputs

• <output>/linkedin-<letter>-<framing>.md per variant

Examples

Builder story brief:

/create-linkedin-content --brief "Built a CRM for my AI agents using just markdown files and Claude Code. No Salesforce, no HubSpot. Works better than both."

→ 3 variants (builder-story, contrarian-insight, workflow-tutorial)

Launch brief:

/create-linkedin-content --brief "Launching goose-aeo — open-source CLI that measures your brand's visibility on AI search engines like ChatGPT, Perplexity, Gemini. npm install, three commands to run."

→ 3 variants (product-launch, problem-first, workflow-tutorial)

Dependencies

• A voice guide at the resolved path
• generate-voice-guide skill (for creating one when missing)

Tips

• Arrow bullets are a LinkedIn signal. When in doubt, use them. They make the post scan-friendly and land as "LinkedIn-native."
• Never open with "I'm humbled" or "Grateful for..." — LinkedIn performative cringe. The voice guide should flag these; honor it.
• Cross-reference ecosystems. LinkedIn readers like knowing the landscape. If you mention a tool, it's fine to note 1–2 comparable ones.
• End with a forward-looking line. "Fast-forward 3 model generations" or "This will change X over the next 12 months" — beats a soft "thoughts?"