Agents are meant to be shared, but existing tooling is not fit for purpose
Summary
The author discusses the difficulty of sharing AI agent workflows across teams and introduces Nairi, a tool for deploying Claude Code-backed agents in Slack with shared access.
Similar Articles
AI agents still suck, so I built my own
The author built a custom AI agent application wrapping Claude Code and upcoming Codex support, focusing on composable workflows and seeking community feedback.
AgentTransfer: OpenSource DropBox for Agents - get agents to share files and talk / work together on projects
AgentTransfer is an open-source tool that allows AI agents to share files and communicate with each other via a lightweight Go binary.
I think I've solved how to seamlessly share your AI agents with coworkers...or anyone for that matter.
The author presents a pattern for sharing AI agents by compiling them into sealed WASM modules that only import a single inference function, allowing others to run the agent using their own local model without API keys or security risks.
There are too many AI agents now and no clean way to showcase what we’ve built
The article highlights the growing problem of managing and showcasing numerous AI agents across scattered platforms, and introduces HiFlixy as a solution for creating organized, shareable agent portfolios.
built 40+ ai workflows for clients this year... 90% of "agents" are gimmicks tbh
The author criticizes the hype around autonomous AI agents, stating they are often ineffective, and recommends building practical, hardcoded AI workflows using tools like n8n and Claude Sonnet for real-world applications.