Open-source AI skills and tools for creators and marketers

2049.news · 23.07.2026, 10:10:01

Open-source AI skills and tools for creators and marketers


A curated list of open-source AI skills and repositories for content creators, marketers and teams looking to automate routine tasks.

Content and social

These tools monitor accounts and topics, extract timely posts to reply to, and generate drafts aligned with a user’s voice and workflow.

  • x-engage — tracks selected accounts and themes, surfaces 5–15 posts requiring immediate replies, and drafts responses in the configured style.
  • head-of-content — monitors X, Instagram, TikTok and YouTube, identifies competitor activity and extracts hooks based on engagement metrics.
  • YouTube Clipper Skill — downloads videos, segments them via FFmpeg, translates and embeds subtitles for creating short-form clips from long videos.
  • Horizon — aggregates Hacker News, Reddit, RSS, Telegram, Twitter and GitHub into personalized digests and automated summaries for faster reading.

Marketing and SEO

Open-source suites provide full audits, customizable interfaces and agentized workflows for search optimization and conversion-focused content production.

  • Agentic SEO Skill — includes 16 sub-skills and 10 specialized agents for end-to-end SEO audits and GEO/AEO workflows compatible with ChatGPT and Perplexity.
  • OpenSEO — offers a user-friendly UI with full source control, allowing teams to extend functionality when off-the-shelf skills are insufficient.
  • marketingskills — a catalog of 47 marketing skills covering CRO, copywriting, SEO and additional growth disciplines for building marketing stacks from scratch.

Client discovery and business fit

Specialized skills analyze public signals, rank prospects by product fit and timing, and surface the most promising outreach targets for early customers.

  • codex-first-customer-finder-skill — accepts a startup URL, assesses public pain signals and demand, and ranks potential first customers by fit and timing; this repository was retweeted by Greg Brockman (president and co-founder of OpenAI).

Agent infrastructure

Frameworks and validators help assemble multi-agent systems, enforce quality checks, and operate agents in parallel from resource-constrained devices.

  • Superpowers — a productivity base for brainstorming, planning and composing subordinate agent skills for iterative workflows.
  • skill-validator — evaluates skills for quality and compliance, counts tokens and detects broken links when many skills accumulate.
  • Orca — tooling for parallel agents, enabling CLI-driven agents such as Claude Code or Codex to run and be controlled from mobile devices.
  • telegram-mcp — a package with 80+ Telegram utilities for reading chats, scheduling messages, managing groups and handling media through a single MCP server.

Video and voice

Open-source audio and video skills support transcription, TTS, voice cloning and agent-level video understanding for content repurposing workflows.

  • CosyVoice — open-source TTS from Alibaba supporting 9 languages, including Russian, able to clone a voice from a short sample with latency from 150 ms.
  • claude-video — enables an agent to "watch" videos by downloading clips, extracting frames and producing transcriptions for downstream processing.
  • notebooklm-py — an unofficial Python API for NotebookLM that functions as both a CLI and a skill integration for agent platforms.

These open-source projects can be combined into pipelines for content production, audience discovery and automation across marketing, product and community workflows.


Related posts

How Claude Cowork Organizes Multi‑Session AI Workflows
Cold Wallet for $0: Turning a Smartphone into Air‑Gapped Storage
Scroll down to load next post