social-distribute

Social Distribute

Generate platform-specific content variants from a blog article and write them to a persistent content file for the automated publishing pipeline. Discord can optionally be posted immediately via webhook after approval. The content file feeds into the directory-driven cron pipeline (content-publisher.sh) for scheduled publishing.

Headless Mode Detection

If $ARGUMENTS contains --headless, set HEADLESS_MODE=true and strip --headless from $ARGUMENTS. The remainder is the blog post path.

Argument format: <blog-post-path> [--headless]

Headless defaults for interactive gates:

• Phase 7 (Discord Approval): auto-selects Skip (never auto-post to external platforms without human approval). channels is always discord, x, bluesky, linkedin-company.
• Phase 9 Step 2 (Overwrite Check): auto-selects Overwrite (content is regenerated from the same blog post, so overwriting is idempotent).

Prerequisites

Before generating content, verify all prerequisites. If a hard prerequisite fails, display the error message and stop. Soft prerequisites display a warning and continue.

1. Brand Guide (hard)

Check if knowledge-base/marketing/brand-guide.md exists.

If missing:

No brand guide found. Run the brand architect agent first to establish brand identity:

Use the brand-architect agent to define our brand.

Stop execution.

2. Blog Post Path (hard)

The skill expects a blog post path as an argument (e.g., /soleur:social-distribute plugins/soleur/docs/blog/my-article.md).

If no path provided or file does not exist:

Provide a path to a blog post markdown file:

/soleur:social-distribute <path-to-blog-post.md>

Stop execution.

3. Discord Webhook URL (soft)

Check if DISCORD_BLOG_WEBHOOK_URL or DISCORD_WEBHOOK_URL environment variable is set.

If both missing:

Neither DISCORD_BLOG_WEBHOOK_URL nor DISCORD_WEBHOOK_URL is set. Discord posting will be skipped (manual output only).

To configure: Server Settings > Integrations > Webhooks > Copy URL from the #blog channel > export DISCORD_BLOG_WEBHOOK_URL="..."

Continue execution -- Discord will be included in the content file's channels field for cron publishing instead of immediate webhook posting.

Content Input

Phase 1: Read Blog Post

1. Read the blog post markdown file at the provided path
2. Parse YAML frontmatter to extract: title, description, date, tags

Phase 2: Gather Current Stats

Run these shell commands to get current component counts:

# Count agents (recursive .md files under agents/)
find plugins/soleur/agents -name '*.md' | wc -l

# Count skills (directories with SKILL.md)
find plugins/soleur/skills -maxdepth 2 -name 'SKILL.md' | wc -l

# Count commands
ls plugins/soleur/commands/*.md | wc -l

# Count departments (non-empty top-level dirs under agents/)
find plugins/soleur/agents -mindepth 1 -maxdepth 1 -type d | wc -l

# Get site URL
cat plugins/soleur/docs/_data/site.json

Extract: agents count, skills count, commands count, departments count, site.url.

Phase 3: Build Article URL

Construct the article URL from site.url and the blog post path:

• Strip plugins/soleur/docs/ prefix from the path
• Replace .md extension with /
• Strip any leading YYYY-MM-DD- date prefix from the filename portion of the path (regex: /\d{4}-\d{2}-\d{2}-(.*)/). Eleventy's page.fileSlug strips this prefix, so URLs must match. If no date prefix exists, leave the path unchanged.
• Prepend site.url

Example: plugins/soleur/docs/blog/2026-03-24-vibe-coding-vs-agentic-engineering.md becomes https://soleur.ai/blog/vibe-coding-vs-agentic-engineering/

UTM Tracking: Derive the campaign slug from the article URL path — strip /blog/ prefix and trailing /. Example: /blog/caas-pillar/ → caas-pillar. If the slug contains characters other than a-z, 0-9, hyphens, or underscores, replace them with hyphens.

Construct platform-specific tracked URLs using this mapping:

| Platform | Tracked URL | |----------|------------| | Discord | <base-url>?utm_source=discord&utm_medium=community&utm_campaign=<slug> | | X/Twitter | <base-url>?utm_source=x&utm_medium=social&utm_campaign=<slug> | | IndieHackers | <base-url>?utm_source=indiehackers&utm_medium=community&utm_campaign=<slug> | | Reddit | <base-url>?utm_source=reddit | | Hacker News | <base-url>?utm_source=hackernews&utm_medium=community&utm_campaign=<slug> | | LinkedIn Personal | <base-url>?utm_source=linkedin-personal&utm_medium=social&utm_campaign=<slug> | | LinkedIn Company Page | <base-url>?utm_source=linkedin-company&utm_medium=social&utm_campaign=<slug> | | Bluesky | <base-url>?utm_source=bluesky&utm_medium=social&utm_campaign=<slug> |

Reddit gets minimal UTM parameters (utm_source only) to reduce spam filter risk — long marketing-looking URLs can trigger irreversible domain reputation damage on Reddit.

Use each platform's tracked URL in the corresponding Phase 5 variant section instead of the bare base URL.

Content Generation

Phase 4: Read Brand Guide

Read the brand guide sections that inform content generation:

1. Read ## Voice -- apply brand voice, tone, do's and don'ts
2. Read ## Channel Notes > ### Discord -- apply Discord-specific guidelines
3. Read ## Channel Notes > ### X/Twitter -- apply X/Twitter-specific guidelines
4. Read ## Channel Notes > ### LinkedIn Personal -- apply LinkedIn personal profile guidelines
5. Read ## Channel Notes > ### LinkedIn Company Page -- apply LinkedIn company page guidelines

If a channel notes section is missing for a platform, generate content using only the ## Voice section.

Phase 5: Generate All Variants

Using the blog post content, stats values, article URL, and brand guide as context, generate all platform-specific variants. The LLM handles template variable substitution (replace {{ stats.agents }} with actual counts), markup stripping (ignore JSON-LD, HTML tags, FAQ accordions), and content adaptation per platform.

Important: Every variant must contain resolved numbers, not template syntax like {{ stats.agents }}. Use the stats gathered in Phase 2.

Lead with the differentiating hook, not the source. For a news-hook / response post (adopting an external term, reacting to an essay or announcement), open EACH variant with the strongest claim about what is unique here — credit the source immediately after, but never bury the original contribution under the news peg. The previews readers actually see — X tweet 1, the LinkedIn first line (shown before "see more"), the Bluesky opener, the Discord first line — must carry the differentiator. Keep that opening hook consistent across all channels for campaign cohesion. Why (#5088): the first loop-engineering draft opened every variant with "Addy Osmani coined…" and buried Soleur's cross-domain expansion in paragraph 3; the operator rewrote all four to lead with the founder-vs-engineer differentiator ("loop engineering replaced the engineer who prompts the coding agent; Soleur replaces the founder who prompts every department").

5.1 Discord Announcement

• Maximum 2000 characters (UTM-tagged URLs are ~60-80 chars longer than bare URLs — account for this in the character budget)
• Include Discord tracked URL
• Match brand voice from ## Voice and ## Channel Notes > ### Discord
• Plain text only (no rich embeds)
• Declarative, concrete, builder-to-builder tone

5.2 X/Twitter Thread

• Hook tweet (standalone value, no "thread" announcement)
• Numbered body tweets (2/ 3/ 4/)
• Final tweet with X/Twitter tracked URL and up to one hashtag
• Each tweet maximum 280 characters
• Match brand voice from ## Voice and ## Channel Notes > ### X/Twitter
• Links only in final tweet
• No emojis in hook tweet

5.3 IndieHackers Building Update

• Markdown format
• Transparent metrics and numbers
• Building-in-public framing
• Include Indi