Content Workflow

Content Workflow

End-to-end content creation pipeline: research, draft, review, distribute.

Install

git clone https://github.com/thatrebeccarae/claude-marketing.git && cp -r claude-marketing/skills/content-workflow ~/.claude/skills/

Pipeline Stages

Stage 1: Research

Gather raw material for the content piece.

1. Topic research — web search, RSS feeds, industry sources
2. Source assessment — credibility, recency, relevance
3. Data extraction — key statistics, quotes, data points
4. Counter-arguments — opposing views, nuance, caveats
5. Angle development — what makes this piece unique

Output: Research brief with sources, data points, and recommended angle.

Stage 2: Draft & Edit

Create and refine the content.

1. Draft — write following the content type framework and brand voice
2. Editorial review against quality criteria:

• Voice consistency
• Structural quality (hook → body → CTA)
• SEO optimization (if applicable)
• Factual accuracy (all claims sourced)
• Readability (appropriate level for audience)

1. Revision — apply specific feedback
2. Final review — verify all issues addressed

Output: Polished draft ready for approval.

Stage 3: Distribute

Adapt the finished piece for cross-platform distribution.

1. LinkedIn post — hook, body, CTA, hashtags
2. Twitter/X thread — numbered tweets, hooks per tweet
3. Email subject lines — 3 options with preview text
4. Pull quotes — 3-5 shareable excerpts for social

Output: Distribution pack with platform-specific adaptations.

Usage Modes

Full Pipeline

/content-workflow [topic]

Runs all 3 stages: research → draft → review → distribute.

Start from Draft

/content-workflow --from-draft [file path]

Skips research. Takes an existing draft through editorial review and distribution.

Distribution Only

/content-workflow --distribute [file path]

Takes a finished, approved piece and creates the social distribution pack.

Content Types

| Type | Research Depth | Draft Length | Distribution | |------|---------------|-------------|--------------| | Blog post | Deep (5+ sources) | 1,500-2,500 words | LinkedIn + Twitter + Email | | LinkedIn post | Light (2-3 sources) | 150-300 words | Twitter adaptation only | | Twitter thread | Light (1-2 sources) | 5-12 tweets | LinkedIn adaptation only | | Newsletter | Medium (3-5 sources) | 500-1,000 words | Twitter + LinkedIn teasers | | Essay | Deep (8+ sources) | 2,000-4,000 words | Full distribution pack |

Status Tracking

Content moves through these statuses:

stub → draft → reviewed → approved → published → measured
                                   ↘ rejected (terminal)

• stub: Topic identified, no content yet
• draft: First draft complete
• reviewed: Editorial review complete
• approved: Human approval gate passed
• published: Live on platform
• measured: Performance data collected
• rejected: Did not pass review, will not publish

Key Principles

1. Research before writing. Even short posts benefit from 10 minutes of source gathering.
2. One piece, many platforms. Write once, adapt to each platform's format.
3. Human approval gate. Never publish without explicit approval.
4. Measure everything. Track performance to inform future content decisions.

For editorial review criteria and platform templates, see [REFERENCE.md](REFERENCE.md).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: thatrebeccarae
• Source: thatrebeccarae/claude-marketing
• License: MIT

Install and usage instructions live in the source repository linked above.