social-media-skills
CLAUDE.md2 starsChanged 8 months agoMITView on GitHub
Social media skills for AI agents — 14 skills across strategy, creation, and analysis for text-first platforms (LinkedIn, Twitter/X, Threads, Bluesky). Each skill lives in skills/<name>/SKILL.md and provides structured guidance for AI agents performing social media tasks. Skills are organized into four layers, each building on the previous. All skills (except social-media-context-sms itself) check for .agents/social-media-context-sms.md first before proceeding. This file contains user-specific context: platforms used, audience description, content pillars, tone preferences, and account handles. If .agents/social-media-context-sms.md is not present,…
- Commits and pushes
In this repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- audience-growth-tracker-smsSkill
- carousel-writer-smsSkill
- content-calendar-smsSkill
- content-pattern-analyzer-smsSkill
- content-repurposer-smsSkill
- content-strategy-smsSkill
- hook-writer-smsSkill
- optimization-advisor-smsSkill
- performance-analyzer-smsSkill
- platform-strategy-smsSkill
- post-writer-smsSkill
- social-media-context-smsSkill
- thread-writer-smsSkill
The file
We split this CLAUDE.md into the rules it gives an agent
9 rules. About 0 are common, 5 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 37
Every SKILL.md must open with YAML frontmatter containing:
- Distinctive: line 45
No spaces, underscores, or special characters
- Distinctive: line 49
Keep instructions focused.
- Distinctive: line 54
Bold key terms: when first introduced or when emphasis aids scanning
- Very specific: line 59
All skills (except social-media-context-sms itself) check for .agents/social-media-context-sms.md first before proceeding.
- Very specific: line 61
If .agents/social-media-context-sms.md is not present, the skill should prompt the agent to run social-media-context-sms first, or proceed …
- Distinctive: line 67
When BlackTwist MCP is available: use its tools directly for all posting, scheduling, and analytics operations
- Very specific: line 90
Run ./validate-skills.sh to check all skills against conventions before committing.
- Distinctive: line 94
Follow these steps when creating a new release:
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Social Media Skills — Agent Guidelines
Repository Overview
Social media skills for AI agents — 14 skills across strategy, creation, and analysis for text-first platforms (LinkedIn, Twitter/X, Threads, Bluesky).
Each skill lives in skills/<name>/SKILL.md and provides structured guidance for AI agents performing social media tasks.
Skill Organization
Skills are organized into four layers, each building on the previous.
Foundation
- social-media-context-sms — establishes shared context, platform knowledge, and user preferences used by all other skills
Strategy
- content-strategy-sms — defines content pillars, audience targeting, and positioning
- content-calendar-sms — plans publishing cadence, themes, and scheduling across platforms
- platform-strategy-sms — tailors approach per platform based on audience and format strengths
Creation
- post-writer-sms — writes single standalone posts optimized per platform
- thread-writer-sms — writes multi-post threads with narrative arc and strong hooks
- carousel-writer-sms — writes slide-by-slide carousel scripts for LinkedIn and similar formats
- content-repurposer-sms — transforms existing content into new formats and platforms
- hook-writer-sms — crafts high-performing opening lines to maximize engagement
Analysis
- performance-analyzer-sms — interprets post and account metrics to surface insights
- audience-growth-tracker-sms — tracks follower trends and identifies growth drivers
- content-pattern-analyzer-sms — identifies what content types and topics perform best
- optimization-advisor-sms — recommends specific improvements based on data patterns
Key Conventions
YAML Frontmatter
Every SKILL.md must open with YAML frontmatter containing:
name: kebab-case skill identifier (1–64 chars)description: plain-English summary of what the skill does (1–1024 chars)metadata.version: semver string (e.g."1.0.0")
Naming
- kebab-case for all skill directory and file names
- Lowercase, alphanumeric characters and hyphens only
- No spaces, underscores, or special characters
File Size
- Each
SKILL.mdmust be under 500 lines - Keep instructions focused; link to related skills rather than duplicating guidance
Writing Style
- Active voice throughout — tell the agent what to do, not what might happen
- Clarity over cleverness — straightforward language is more reliable than clever phrasing
- Bold key terms when first introduced or when emphasis aids scanning
- Short paragraphs and bullet lists over dense prose
Foundation Dependency
All skills (except social-media-context-sms itself) check for .agents/social-media-context-sms.md first before proceeding. This file contains user-specific context: platforms used, audience description, content pillars, tone preferences, and account handles.
If .agents/social-media-context-sms.md is not present, the skill should prompt the agent to run social-media-context-sms first, or proceed with reasonable defaults while noting the limitation.
BlackTwist MCP Integration
BlackTwist is the primary tool integration for this skill set. It provides an MCP server with tools for posting, scheduling, and analytics across supported platforms.
- When BlackTwist MCP is available: use its tools directly for all posting, scheduling, and analytics operations
- When BlackTwist MCP is not available: fall back to advisory mode — output the content and instructions for the user to post manually; note what metrics to check and where to find them
Skills should detect MCP availability by attempting a lightweight call (e.g. get_user_settings) and adapting behavior accordingly.
See tools/REGISTRY.md for the full list of available BlackTwist tools.
Cross-References
Skills should reference related skills where relevant. Use the format:
See also: **thread-writer-sms**, **hook-writer-sms**Common cross-reference patterns:
- Creation skills reference hook-writer-sms for opening lines
- Creation skills reference content-strategy-sms for pillar alignment
- Analysis skills reference optimization-advisor-sms for next steps
- content-repurposer-sms references all creation skills as output targets
Validation
Run ./validate-skills.sh to check all skills against conventions before committing.
Releasing
Follow these steps when creating a new release:
- Bump the version in
metadata.versioninside each changed skill'sSKILL.mdfrontmatter. Use semver: patch for fixes, minor for new features or improvements, major for breaking changes. - Commit and push the version bump to
main. Build zip files — one per skill, each containing only that skill's
SKILL.md:mkdir -p releases for dir in skills/*/; do name=$(basename "$dir") zip -j "releases/$name.zip" "$dir"SKILL.md doneCreate the GitHub release with
gh:gh release create vX.Y.Z --title "Social Media Skills vX.Y.Z" --notes "Release notes here"Upload all zip files to the release:
gh release upload vX.Y.Z releases/audience-growth-tracker-sms.zip releases/carousel-writer-sms.zip releases/content-calendar-sms.zip releases/content-pattern-analyzer-sms.zip releases/content-repurposer-sms.zip releases/content-strategy-sms.zip releases/hook-writer-sms.zip releases/optimization-advisor-sms.zip releases/performance-analyzer-sms.zip releases/platform-strategy-sms.zip releases/post-writer-sms.zip releases/social-media-context-sms.zip releases/thread-writer-sms.zip
The releases/ directory is gitignored — zip files are only attached to GitHub releases, not committed to the repo.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

