Content Repurposing

Outcome

One piece of source content turned into 10-16 platform-native assets, saved to projects/mkt-content-repurposing/{YYYY-MM-DD}_{campaign-name}/. Each file is ready to copy-paste and publish — not a reworded version of the same post, but content that feels like it belongs on each platform.

Context Needs

Load what exists. Skip what doesn't. This skill works standalone.

Dependencies

Step 1: Check What We're Working With

Load brand context per the table above. Tell the user what's available:

Brand context:
  Voice Profile     [loaded / not found]
  Samples           [loaded / not found]
  Learnings         [N entries / none yet]

If voice profile exists, note the tone in one line. If not, work with a neutral professional voice and mention that /mkt-brand-voice would make the output sharper.

Step 2: Get the Source Content

Take whatever the user gives you — pasted text, a URL, a file path. Valid sources:

If they give a URL, try WebFetch first. If that fails (JS-heavy site, bot protection), fall back to firecrawl-scraper (needs FIRECRAWL_API_KEY in .env). If both fail, ask them to paste the content.

Step 3: Break It Into Atoms

Pull the source apart into the pieces you'll remix across platforms:

1. The core insight — if someone remembers one thing, this is it
2. Supporting points — 3-7 ideas that build the case
3. Stories and examples — specific, concrete moments
4. Numbers and proof — data, stats, results
5. Spicy takes — opinions that challenge what people assume
6. Action steps — things the reader can actually go and do
7. Quotable lines — punchy phrases that work as standalone posts

Show the atoms to the user before generating. "Here's what I pulled out. Anything to add or drop?"

Step 4: Check for Algorithm Changes

Before writing anything, do a quick search for each target platform:

"{platform} algorithm update {current month} {current year}"

Compare what you find against the platform references. If something meaningful changed, flag it. If search isn't available, use the reference data and note it.

Step 5: Pick the Platforms

Ask: "All 8 platforms, or specific ones?"

Default to all unless they narrow it. If you have ICP data, recommend the platforms where their audience actually hangs out.

Step 6: Adapt the Voice

Read references/voice-adaptation.md for the full adaptation framework.

The same insight needs to hit differently on each platform. Take the user's voice profile and adjust the register, pace, length, and vocabulary to match where it's going. Same person, different room.

Step 7: Write the Content

For each platform, read the relevant reference:

• LinkedIn + Twitter/X: references/platform-text.md
• Instagram + TikTok + YouTube: references/platform-visual-video.md
• Threads + Bluesky + Reddit: references/platform-emerging.md

Use the hooks, templates, format specs, and CTAs from those references. Write content that feels native to each platform — not the same post reformatted.

For each piece, include:

• The content itself (publish-ready)
• Format recommendation (carousel, thread, reel script, etc.)
• Posting time suggestion (if learnings have data)
• A CTA that fits the platform

Step 8: Content Calendar (If Asked)

If they want a weekly schedule, read references/voice-adaptation.md for the 5-3-2 weekly rhythm and map the assets across the week.

Step 9: Humanizer Gate

Before saving, run all generated content through tool-humanizer in pipeline mode. Use deep mode if brand_context/voice-profile.md was loaded, standard otherwise. Read tool-humanizer/SKILL.md and apply its pattern detection + replacement process to each platform's content. Only show the score summary if the delta exceeds 2 points.

Step 10: Save Everything

Create: projects/mkt-content-repurposing/{YYYY-MM-DD}_{campaign-name}/

One file per platform: linkedin.md, twitter-x.md, instagram.md, tiktok.md, youtube.md, threads.md, bluesky.md, reddit.md. Add calendar.md if they asked for scheduling.

Show actual content previews — the hook and format for each platform. Don't just list file paths.

Rules

Updated when the user flags issues. Read before every run.

Self-Update

If the user flags an issue with the output, add a dated entry under ## Rules in this skill's SKILL.local.md (create it if missing) and log it to context/learnings.md under ## mkt-content-repurposing. Write the general rule the correction implies, not the one-off fix.

Troubleshooting

Source is too thin: Under 300 words doesn't give enough to work with. Ask for more or suggest combining two pieces. Output sounds generic: Probably missing brand context. Offer /mkt-brand-voice or ask for 2-3 sentences in their voice as a quick reference. Platform doesn't fit: Don't force it. Skip platforms where their audience doesn't exist unless they specifically ask. Algorithm data might be outdated: If the references are 3+ months old and search turned up nothing, say so and proceed with best available. They want to schedule posts: This skill writes the content. Point them to Buffer or Hootsuite for scheduling — include optimal posting times in the output so they can plug it straight in.