alirezarezvani/claude-skills +392 ★ this day + Follow

What it actually is

This is a curated content repository, not an application: it packages 337+ "skill" definitions (SKILL.md instruction files plus supporting Python scripts and reference docs) that give AI coding agents — Claude Code, OpenAI Codex, Gemini CLI, Cursor, and others — domain-specific workflows for engineering, marketing, compliance, C-level advisory, research and productivity tasks. Each skill folder (e.g. engineering-team/skills/senior-computer-vision, c-level-advisor/skills/ciso-advisor) bundles a Markdown instruction file with stdlib-only Python CLI tools that the agent can invoke. It targets developers and teams who want off-the-shelf, reusable expertise packages rather than writing agent prompts from scratch.

Why it's growing

The 392-star single-day jump aligns with a fast release cadence: v2.8.0–v2.9.0 landed within roughly a week (19–28 May), each adding new domains (business-operations, commercial, enterprise Research Operations) and fixing regressions (e.g. plugin.json). Commit history shows near-daily feature merges — deep-research, zero-hallucination-coder, agent-decision-receipts, local-seo-manager — each going through multiple automated-review fix cycles before merge, suggesting active promotion and iteration rather than a one-off launch. The README's explicit "13 platforms, 338 skills" framing and badges (Skills, Agents, Commands counts) also read as designed for shareability, which plausibly drives discovery traffic, though the graph facts alone can't confirm the exact cause of the spike.