Content Repurposer: 1 Memo, 5 Outputs

The Solution

Content Repurposer is a self-contained AI agent skill that takes one source — a memo, outline, transcript, or rough notes — and produces five ready-to-publish deliverables in a single automated run.

You are not asking the AI to "write more content." You are asking it to restructure the same core message for each channel's rules. One source of truth. Five angles. Seven output files.

What you get per channel

Every output follows bundled media defaults: length limits, structure templates, hook patterns, and CTA placement — so you spend time reviewing, not rebuilding from zero.

Why Content Repurposer

Works in any language you specify.
Set {LANGUAGE} to en, ja, or any code you need. The logic is universal; the output language is yours.

Agent-agnostic.
Runs on Claude Code, Codex, Cursor, AntiGravity, MANAS, and any AI coding agent that can read markdown skill files and write local files. No vendor lock-in.

Zero dependencies.
No API keys. No plugins. No paid integrations. No database. No GPU. The skill is plain markdown — upload it and run.

Anti-hallucination built in.
The skill will not invent statistics, case studies, testimonials, or facts not present in your source. Missing information is flagged as <<CONFIRM: …>> instead of filled with placeholders or fabricated content.

Optional voice profile.
Copy voice-profile_TEMPLATE.md to voice-profile.md, fill in your tone, audience, banned phrases, and a sample paragraph. All five channels inherit your brand voice automatically.

Self-improving (optional).
When you give correction feedback ("X posts were too long", "newsletter tone too formal"), the skill can append rules to a local override file and reference them on future runs.

Who It's For

• Solo creators publishing across blog, email, X, YouTube, and Instagram
• Marketing teams running content-led growth with limited headcount
• Ghostwriters and agencies batching client deliverables from one brief
• Anyone who already has ideas but hates reformatting the same message five different ways

If you publish on two or more channels and feel the "rewrite tax" every week, this skill pays for itself on the first run.

Setup (About 2 Minutes)

1. Download SKILL.md and voice-profile_TEMPLATE.md from your purchase.
2. Place SKILL.md in your agent's skills folder:

• Cursor: .cursor/skills/content-repurposer/SKILL.md
• Claude Code: .claude/skills/content-repurposer/SKILL.md
• Codex: .codex/skills/content-repurposer/SKILL.md (or your environment's skills directory)

3. Optional — voice profile: Copy the template to voice-profile.md. Fill in Sections 1–3 (identity, audience, tone), Section 6 (CTA defaults), and Section 8 (sample paragraph) for best results.
4. Run with your source material, a one-sentence core message, and an output folder.

Default output directory: ./output/content-repurposer/ (customizable).

Sample Prompt

Repurpose this memo into all five channels.
Core message: Batching one idea across five channels beats writing five separate pieces from scratch.
Audience: solo creators and small marketing teams.
Language: en
Output folder: ./output/this-week/
Voice profile: ./voice-profile.md

[paste your source memo here]

The agent runs seven phases internally and writes all deliverables to files. Chat receives only a completion summary — not a wall of generated text.

Files You Receive Per Run

Suggested publish order: blog first → newsletter → X thread (posts 01–03) → YouTube → Instagram.

How It Works (Pipeline Overview)

1. Extract — Pull theme, core message, audience, key points, and gaps from your source (internal step).
2. Master outline — Map channel-specific angles before writing.
3. Blog — Full article with YAML frontmatter and headline variants.
4. X posts — 10 posts using a fixed angle table (contrarian, question, pain-reframe, proof, CTA, etc.).
5. Newsletter — Hook → insights → single primary CTA.
6. YouTube — 5–10 minute read-aloud script with [MM:SS] timestamps.
7. Instagram — 10 slides from cover hook to CTA + handle.
8. Completion report — Summary, <<CONFIRM>> items, and per-channel pass/fail check.

You can also request a subset of channels (e.g., blog + newsletter only). Unrequested phases are skipped and noted in the report.

Requirements

You need:

• Any AI coding agent that reads markdown skill files and writes local files
• A source memo, outline, or transcript (text)
• A one-sentence core message (what the reader should believe or do)

You do not need:

• API keys or cloud LLM subscriptions beyond what your agent already uses
• External plugins, databases, or image-generation services
• Personal data, vault paths, or project-specific configuration baked into the skill

Instagram output is text + design-tool prompt up to the carousel copy. Image generation is intentionally out of scope — use Canva, Manus, or your preferred design tool with the included export prompt.

What's Not Included (By Design)

• Personal profiles, fixed vault paths, or author-specific style seeds
• Automatic image or video generation
• API integrations with third-party publishing platforms

These boundaries keep the skill portable, privacy-safe, and runnable on any machine.

Version

v1.0.1 — Includes bundled voice-profile_TEMPLATE.md and voice-profile parsing rules across all five channels.

Support

After purchase, refer to the included install guide and sample test memo for a quick validation run. If you hit <<CONFIRM: …>> flags in your output, those mark gaps in your source — fill them in and rerun the affected channel.

Content Repurposer — one memo, five channels, one run.