One routine for every agent
A five-person team gives Claude Code and Codex the same plan, test, verify and review routine.
Give every coding agent on your team the same discipline: plan, test, debug, verify, review and commit.
Eleven self-contained plugins that coding agents pick up on their own. The workflow skills make agents write the failing test first, find the root cause before fixing and verify before calling anything done. Around them sit the tools the same teams use every day: Gemini batch jobs, video analysis, Google Workspace and Cloudflare. We curate and harden the strongest community skills, add our own, and ship the set as one marketplace.
Builds on skills from obra/superpowers, pchalasani/claude-code-tools and Anthropic’s official plugins, all MIT.
AI makes code cheap, so the discipline a team gives its agents decides the quality of what ships.
We wrote ours down as skills, commands and hooks while delivering client work, then split them into plugins that install one at a time. Every project sharpens the playbook, and the next project starts from the sharper version.
A five-person team gives Claude Code and Codex the same plan, test, verify and review routine.
A consultancy installs the pack on the first day of a project and adds a client-specific plugin on top.
Batch-analyse competitor demo videos, sort Drive and Gmail, and run Gemini jobs from the same agent session.
git clone https://github.com/iofold/ainative-claude-plugins.git
cd ainative-claude-plugins
chmod +x setup.sh
./setup.sh # --skip-optional skips non-essential componentsFull documentation is in the repository README.
Tell us what the work is, what data it touches and what constraints apply. We’ll work out what to build and how to measure it.