Marketing Team

---
name: marketing-team
description: "Content repurposing system — takes source content and generates platform-specific versions for ALL channels, with batch review and sequential publishing. Use when: repurposing content across platforms. Triggers on: marketing team, repurpose content, distribute content, content across platforms, publish everywhere."
---

# Marketing Team — Content Repurposing Orchestrator

Turn one piece of content into platform-specific versions for every channel, review as a batch, publish everywhere.

## Prerequisites

- Playwright MCP for publishing (LinkedIn, Substack, Skool, Reddit)
- X API via `{{X_API_SCRIPT_PATH}}` for Twitter/X
- If Playwright unavailable: content saves as markdown for manual publishing

---

## Workflow

**IMPORTANT: Lazy-load files. Only read files when the step that needs them is reached. Do NOT read everything upfront.**

### Step 1: Content Discovery

Read ONLY: `{{PROFILE_PATH}}`

Ask: **"Do you already have content to repurpose, or want me to find something trending in your niche?"**

**A) Find me something:**

**Phase A - Trend Detection (parallel):**
1. Google Trends scrape via `mcp__crawl4ai-mcp__scrape`:
   - `https://trends.google.com/trending?geo=US&category=12`
   - `https://trends.google.com/trends/explore?q=AI%20SDR,AI%20sales,cold%20email%20AI,outbound%20automation&date=now%207-d`
   - If scrape fails, fall back to Phase B
2. Autocomplete via `mcp__serper__google_search_autocomplete`: "AI SDR", "AI cold email", "outbound automation", "AI sales agency"
3. News velocity via `mcp__serper__google_search_news` (last 48h): "AI sales", "AI outbound", "AI SDR", "cold email AI"

Compile 3-5 trending themes.

**Phase B - Content Search (parallel, informed by trends):**
- `mcp__serper__google_search_videos`: trending keywords + niche terms
- `mcp__serper__google_search`: site:linkedin.com/substack.com/reddit.com + trending keywords
- `mcp__serper__google_search_news`: broader niche (last 7 days)
- `mcp__exa__web_search_exa`: semantic search for newsletters, Skool, podcasts
- `mcp__serper__google_search_scholar`: only if data-heavy angle exists

**Phase C - Present:**
```
TRENDING TOPICS DETECTED:
- [keyword] - [source: Trends / autocomplete / news velocity]

CONTENT TO REPURPOSE:

TRENDING:
1. [Platform] "Title" by Creator - stats - WHY: [angle] - TREND: [which]
2. ...

UNDER THE RADAR:
4. [Platform] "Title" - WHY: [angle]
5. ...
```

Ask: **"Pick one (or multiple) to repurpose, or tell me a different topic."**

Scrape picked content, save as `source-content.md` in topic folder.

**B) I have content:** User provides file path, URL, or pastes directly.

**C) I have a topic:** Narrow discovery search to that topic.

### Step 2: Research Gate (Optional)

Ask: **"Deep research report, or just repurpose the source directly?"**

- **Deep research**: follow `~/.claude/skills/youtube-research/SKILL.md`, save `research.md`
- **Just repurpose**: use source content directly
- **Has report**: list folders in `{{CONTENT_DIR}}/`, read selected `research.md`

Then ask:
1. "Any personal experience or story?"
2. "Any platforms to skip?"
3. "YouTube script too?"

### Step 3: Extract Core Message

Distill thesis, 3-5 key points, best story, best data point. Display and get user approval before proceeding.

### Step 4: Spawn Generation Agents

NOW read: `~/.claude/skills/agent-team/teams/marketing-team.md` (team template with agent playbooks)

Use TeamCreate, spawn 4 agents per the template. **Pass file PATHS, not contents:**
- Paths for agents to read: source content, `{{PROFILE_PATH}}`, `~/.claude/skills/marketing-team/references/platform-rules.md`
- Inline: core message, personal story, topic folder path, skip/opt-in flags

Save all output to: `{{CONTENT_DIR}}/[topic-slug]/`

### Step 5: Audit

After all 4 agents finish, spawn auditor per team template. Pass topic folder path + file paths.

### Step 6: Batch Review

Read `audit-report.md`. Do NOT display full content. Output only:
```
Content generation complete.
Audit: [X passed / Y auto-fixed / Z flagged]
All files at: {{CONTENT_DIR}}/[topic-slug]/
Review the files, then tell me: approve all, reject, or edits needed?
```

Single approval triggers publishing.

### Step 7: Visual Selection

**LinkedIn** - for each post ask: "Personal photo or branded diagram?"
- Photo: list from `{{PHOTOS_DIR}}/`
- Diagram: invoke `drawio` skill (1080x1350 PNG)

**Twitter** (optional): invoke `drawio` skill (1200x675 PNG)

### Step 8: Sequential Publishing

NOW read: `~/.claude/skills/marketing-team/references/publishing-playbook.md`

Publish in order:
1. **Blog** - write to {{PROJECT_NAME}}, git commit + push
2. **LinkedIn** - Playwright, schedule consecutive days {{PREFERRED_POST_TIME}}, attach photos
3. **Substack** - Playwright, save as draft by default
4. **Twitter** - ask "now or schedule?" Post via `{{X_API_SCRIPT_PATH}}`
5. **Skool** - Playwright
6. **Reddit** - Playwright

Rules: never stop on failure, log error + continue. Report success/failure per platform.

### Step 9: Publish Log

Save `publish-log.md` to topic folder with platform statuses, errors, and generated file list.