Content Dispatch Hub — One Topic, Every Platform

Distribute your content to multiple platforms from a single topic. Each platform skill independently adapts content for its target audience — developers on Dev.to, newsletter readers on Ghost, and more.

Get YouMind API Key → · More Skills →

Onboarding

⚠️ MANDATORY: When the user has just installed this skill, run the onboarding flow IMMEDIATELY. Translate to the user's language.

The onboarding flow sets up TWO things in ONE conversation: platform roster (where to publish) and author DNA (how to write). These are NOT separate steps — they happen together during the first dispatch.

Phase 1: Platform roster (30 seconds)

✅ Content Dispatch Hub installed!

Quick setup — which platforms do you use?

1. Dev.to 2. Hashnode 3. WordPress 4. Tumblr 5. Ghost
2. Beehiiv 7. Kit 8. X/Twitter 9. WeChat (公众号) 10. Qiita

(list numbers, e.g. "1, 5, 8")

After user selects:

1. Save to ~/.youmind/dispatch-roster.yaml → active_platforms
2. For each selected, check skill installed + YouMind connection via validate
3. Report status table (✅ connected / ❌ → setup link)

Phase 2: Author DNA bootstrap (integrated into first dispatch)

Do NOT ask the user to fill a profile form. Instead, bootstrap the author profile from one of two paths:

Path A — User has articles in YouMind KB (preferred):

1. Call YouMind search API to fetch the user's recent 10–15 articles
2. Analyze across articles: avg paragraph length, code density, language distribution, opening patterns, depth level, typical word count, tone markers
3. Propose a profile:

📝 I analyzed your recent articles. Here's what I see:

• Voice: Conversational-technical, first-person practitioner
• Audience: Mid-level developers (you assume REST/Git knowledge)
• Code density: High (avg 4 code blocks per article)
• Depth: Practitioner-level (how-to with real gotchas, not beginner intros)
• Languages: English primary, Chinese secondary
• Signature moves: You always lead with the problem, include a gotchas section, and end with working code

Sound right? I'll use this as your baseline across all platforms. You can adjust anytime.

4. User confirms/adjusts → save to ~/.youmind/author-profile.yaml
5. Immediately proceed to first dispatch with the profile active

Path B — New user, no articles (cold start): Embed 3 targeted questions INTO the first dispatch request, not as a separate quiz:

You said: "帮我写一篇关于 AI agents 的文章发到 Dev.to 和 X"

Great! Since this is your first dispatch, 3 quick questions so I write in YOUR voice:

1. 这篇文章的风格？ (a) 代码先行，先看效果再讲原理 (b) 故事驱动，用经历带出技术点 (c) 数据说话，benchmark 和对比为主 (d) 手把手教学，假设读者是入门者

2. 你的读者是谁？ (a) 做产品的开发者 (b) 正在学习的学生/新人 (c) 做技术决策的 TL/架构师 (d) 泛科技读者

3. 你的内容里绝对不要出现什么？ (自由回答 — 例如"不要营销话术"、"不要 AI 味"、"不要太长")

After user answers:

1. Map answers to author-profile.yaml fields
2. Save profile
3. Immediately continue the dispatch — no second round-trip. The first article is generated WITH the profile already active.
4. After publishing, confirm: "Saved as your writing profile. Next time just say '帮我分发', I know your style."

Why this matters

Roster + Profile updates after onboarding

"Add Ghost to my platforms" → update ~/.youmind/dispatch-roster.yaml "Remove X/Twitter" → update roster "I want to write more beginner content" → update ~/.youmind/author-profile.yaml "Analyze my recent articles again" → re-run Path A bootstrap

Usage Modes

Mode 1: Single Platform Routing

User specifies one platform → dispatch directly invokes the corresponding platform skill.

"Write about AI agents for Dev.to"         → invoke youmind-devto-article
"Write a weekly issue for Beehiiv"         → invoke youmind-beehiiv-article
"Write a Tumblr post about indie hacking"  → invoke youmind-tumblr-article
"Draft a public creator post for Kit"      → invoke youmind-kit-article
"Post a thread about startup culture on X" → invoke youmind-x-article
"发布一篇技术文章到 WordPress"                → invoke youmind-wordpress-article

Mode 2: Multi-Platform Dispatch

User specifies multiple platforms → dispatch invokes each platform skill sequentially with the same topic.

"Write about AI agents and publish to Dev.to, Beehiiv, and Kit"
"把这个话题发布到 Dev.to 和 Hashnode"

Mode 3: All-Platform Broadcast

User wants maximum reach → dispatch to all installed platform skills.

"Publish everywhere about the future of AI coding"
"全平台发布：AI 编程的未来"

Platform Skills Are Independent

Each youmind-{platform}-article skill is independently installed, independently published, and fully self-contained. They work without this dispatch hub. Dispatch is an optional orchestration layer — not a dependency, not a framework that platform skills must fit into.

How dispatch connects to platform skills (when both are installed)

• Discovery, not coupling. Dispatch scans for installed youmind-*-article skills at runtime. It optionally reads a small dispatch-capabilities.yaml at each skill's root to learn what operations that skill supports (generate, adapt, condense, translate, etc.) and its hard limits. If a platform skill does not ship this file, dispatch falls back to sensible defaults and still invokes it normally.
• Brief is an optional extra, not a requirement. When dispatch invokes a platform skill, it passes a content brief (topic, angle, keywords, optional resolved_author block). Platform skills that understand these fields use them as richer context; skills that don't simply ignore the extras and work as usual.
• Author profile lives in the user's home directory, not in this skill. Both this hub and platform skills read ~/.youmind/author-profile.yaml — the canonical shared location defined in shared/YOUMIND_HOME.md. Uninstalling dispatch does not remove the profile; platform skills continue to work standalone with full DNA support.
• Results flow back as a standardized shape (status, URL, result_links, title, conformance_report) so dispatch can aggregate multi-platform runs. Each platform skill publishes this same result shape regardless of whether dispatch is present — it's a clean output API, not a contract with dispatch.

Zero coupling obligations for platform skills

• Platform skills do NOT import from dispatch.
• Platform skills do NOT require dispatch to be installed.
• Platform skills do NOT break if dispatch is uninstalled.
• Platform skills do NOT need to be updated when dispatch changes.
• dispatch-capabilities.yaml is opt-in metadata that lets dispatch route more intelligently. Removing it reverts to defaults; it never breaks the platform skill.

The optional integration protocol is documented at shared/DISPATCH_CONTRACT.md (v1.0). Platform skills that want richer dispatch integration may follow it; those that don't, don't.

Platform Registry

Auth: All platforms use YouMind OpenAPI. You need only a youmind.api_key — platform credentials are stored encrypted in YouMind Connector Settings. No local platform keys.

Note: Each platform skill must be installed separately. Dispatch checks your roster (dispatch-roster.yaml) on each run.

Quick Skill Hints

Use these as lightweight routing clues before you open the full platform skill:

• youmind-devto-article: code-first Markdown posts, TL;DR, practical developer tutorial voice
• youmind-hashnode-article: deeper developer-blog storytelling, series-friendly, canonical-aware cross-posting
• youmind-wordpress-article: HTML CMS publishing, taxonomy/excerpt aware, company-blog friendly
• youmind-tumblr-article: voice-driven creator posting, one title plus one clean rich-text body
• youmind-ghost-article: editorial publication tone, excerpt/tag aware, newsletter-style HTML
• youmind-beehiiv-article: newsletter-first writing, subject/preview/web-email surfaces matter
• youmind-kit-article: creator broadcast format, inbox/feed readability, compact CTA structure
• youmind-x-article: concise hooks, thread splitting, sparse hashtags, native social adaptation
• youmind-wechat-article: WeChat-native rich formatting, draft-box publishing, built-in direct-send for ready articles
• youmind-qiita-article: Japanese developer Markdown, environment/setup context, note-box friendly

Dispatch Pipeline

Step 1: Parse Request + Load Roster

• Extract topic/brief from user input
• Load ~/.youmind/dispatch-roster.yaml for the user's active platform list
• Resolve target platforms:
  • User says "帮我分发" / "publish everywhere" → use active_platforms from roster
  • User specifies platforms explicitly → use those (even if not in roster)
  • No roster exists → ask user which platforms (triggers onboarding)
• For each target platform: verify skill installed + connection status from roster
• Skip platforms that are not installed or not connected (with clear message + setup link)

Step 2: Load Author Profile + Generate Content Brief

2a — Load author profile (if exists): Read ~/.youmind/author-profile.yaml (canonical shared location — see shared/YOUMIND_HOME.md). This file captures the user's cross-platform writing DNA — voice, audience, content preferences, and per-platform overrides. See references/author-profile-spec.md for format. Fallback: if the ~/.youmind/ path is missing but skills/youmind-article-dispatch/author-profile.yaml exists (legacy), offer to migrate.

If no profile exists → skip this step; dispatch works without it (just less personalized). After the first successful dispatch, offer to help the user build one.

2b — Generate content brief (see references/content-brief-format.md):

topic: "The core topic sentence"
angle: "The specific angle or atomic insight"
keywords: ["keyword1", "keyword2", "keyword3"]
language: "en"        # or zh, ja, etc.
source_material: []   # from YouMind knowledge mining (if available)
resolved_author: {}   # per-platform resolved profile (computed in Step 3)
constraints: {}       # any user-specified constraints

YouMind knowledge mining: If the user has a YouMind API key configured in any platform skill's config, run mine-topics ONCE and include the results in the brief. This avoids redundant YouMind calls across platform skills.

2c — Resolve author profile per platform: For EACH target platform, merge the author profile with that platform's DNA using the rules in references/content-adaptation-matrix.md. This produces a resolved per-platform profile — NOT the raw author file. Each platform skill receives only the resolved, relevant fields for its platform. This prevents raw-profile leakage and irrelevant context bleed.

If a depth mismatch is detected (e.g., author targets "expert" but platform audience is "general"), the merge triggers a mandatory outline adaptation step — not just a warning. See matrix Priority 5.

Step 3: Dispatch to Platform Skills

For each target platform, invoke the platform skill with the content brief as context.

Sequential mode (default): Run each platform skill one at a time. The agent reads each skill's SKILL.md and follows its pipeline, passing the content brief as the initial topic/context.

Parallel mode (if subagents available): Spawn one subagent per platform. Each subagent independently runs its platform skill's full pipeline. Use this when the agent platform supports parallel subagents.

MANDATORY: Pass the content brief to each platform skill. When invoking a platform skill, include:

1. The topic and angle from the brief
2. Any YouMind knowledge context already mined
3. The instruction to adapt content for that specific platform's audience

Step 4: Collect Results

Gather results from all platform dispatches:

Step 5: Summary Report

Present a summary table to the user:

## Dispatch Results

| Platform | Status | Title | URL |
|----------|--------|-------|-----|
| Dev.to | ✅ Published | "AI Agents: A Practical Guide" | https://dev.to/... |
| Ghost | ✅ Published | "Why AI Agents Matter for..." | https://newsletter.example.com/... |
| X/Twitter | ✅ Thread posted | "🧵 AI Agents are changing..." | https://x.com/... |

After the table, add a compact Result links section for each platform so the user can click straight into the post page, dashboard, or stats entry point. If a platform has no direct analytics URL, include the best platform entry URL instead.

Resilience Rules

References