Descriptions:
Nick Saraev delivers a full course on building Claude “skills” — reusable, self-patching agent workflows — for business applications rather than personal assistant use cases. Drawing from his experience managing a business he reports at over $4 million in annual profit primarily through AI agents, Saraev demonstrates five skills live inside the anti-gravity interface running multiple Claude Code plugins simultaneously.
The featured skills include automated CRM follow-up that reads email chains and generates personalized outreach matched to each lead’s pipeline stage, lead scraping workflows, and content generation pipelines. The key technical concept is that skills are markdown files with a specific structure: once an SOP or task list is fed to Claude with the skill specification, it generates the skill definition automatically. Critically, skills self-patch — if the agent encounters an error or missing resource during execution, it diagnoses and fixes the skill file before continuing.
Saraev notes that the skills format is now supported across all major AI coding platforms: Anthropic’s Claude Code, OpenAI’s Codex, and Google’s Gemini CLI all use similar specifications, meaning skills built for one platform transfer with minimal adjustment. The course covers the full lifecycle from generating a skill definition from a bullet-point SOP, to first-run testing, to ongoing maintenance. For practitioners running service businesses, e-commerce operations, or consulting practices, the video provides a practical framework for automating knowledge work that would otherwise require multiple full-time employees.
📺 Source: Nick Saraev · Published March 02, 2026
🏷️ Format: Course Lesson
