skills

Why we included this project

If you're already using Claude Code, this repo is a handy way to get agents that trigger themselves, pick up tasks, run and verify their own work, and log what they learned, so effort compounds without you prompting every step. The two flagship skill sets are a codebase harness that makes any repository agent-ready with run/test/verify/ship steps and isolated cloud boxes per agent, and a loops skill that shares a file-based knowledge base across agents. Beyond those, standalone utilities cover auditing agent context against Anthropic's guidance, delegating to any CLI agent in detached tmux sessions, and an SEO growth playbook that can run on a schedule. It's an opinionated collection of patterns the maintainers run in production, so it's a good starting point if you want to adopt setups that have been exercised in real use rather than build your own from scratch.