Building the agent-native toolchain.
Open source CLIs designed for a world where AI agents ship your code.
| Project | What it does | Install |
|---|---|---|
| agentic-dna | Governance and knowledge system for AI agent fleets. | npm i -g agentic-dna |
| openclaw-extensions | CLI utilities for OpenClaw β cron frequency, adaptive scheduling. | npm i -g openclaw-extensions |
| agent-git | Git wrapper that tracks cloned repos and prevents duplicate clones. | brew install exisz/tap/agent-git |
| avercel | Vercel CLI wrapper β stops trailing newlines, blocks vercel deploy, audits env vars. Built-in AgentSkill. |
npm i -g avercel |
| lazyjira | Zero-dependency CLI for Jira Cloud. Like lazygit, but for Jira. | pip install lazyjira |
| agentbase | Agentic Database that supports markdown and trello | npm i -g @exisz/agentbase |
| cronology | Ephemeral task watcher service. | npm i -g cronology |
| agentrealm | Agentless sandbox workspaces for AI agents. Bundles realm context into one prompt. | npm i -g agentrealm |
| clawdius | Another claw-like. | npm i -g clawdius |
- Agent-first β AI agents are writing and deploying code now. They hit the same bugs humans do, but can't debug them. My tools catch the bugs before they happen β guardrails baked into the CLI, not the prompt.
- Wrap, don't replace β extend tools you already use.
avercelwrapsvercel,lazyjirawraps Jira's API. Zero learning curve. - Ship with a skill β every tool includes an AgentSkill so AI agents know how to use it out of the box.
- Zero dependencies when possible β less bloat, more reliability
- CLI first β if it can't be done from a terminal, it's not done
If my tools saved you (or your agent) time, consider giving them a β




