@earendilworks Beeing honest, I'm not sure how to feel about this. Seems a good product but I'm worried about how much this could affect the future core code in Pi, maybe mixing tech and business decisions. I trust you guys, bu it's an open gate now.
@jorgeucano Orbstack funciona joya, totalmente swipeable como engine para Docker, Colima también lo usé, parece tener mejor rendimiento pero no es tan amigable, cambiar el engine para Docker es solo un comando así que podés probar los 3 muy fácil
@fjpedrosa86@mattpocockuk@unclebobmartin Yo me basé en esto hace un tiempo y tengo algo implementado con sub agentes en herdr + pi que me gusta llamar GDD (gates driven develooment) 😁
Hey @maxedapps , maybe you’ll be interested in this… my own recipe for pi-herdr sub-agents. Claude included.
No support si provided, haha.
https://t.co/WubH79lsX6
@maxedapps Built a rough Herdr-based Pi extension for multi-agent workflows: markdown agents, tools (spawn/ger_status/give_feedback/close), and a skill. Works, but:
1) I'm polling statuses... Any events based solution?
2) default Git worktrees or only when needed (research don't need wt)?
@RalphUtz@IndraVahan I'm trying this, but it's slow and fails a lot with tool calling and debugging bugs, as an example, was not able to fix a bad Tailwind configuration for tanstack-start. What are your config/resources? I'm on Q4_K_M - 16.5 GB on 24 GB RAM.
@tannerlinsley Are there plans to offer “official” TanStack skills—similar to what Vercel did? I think that would be really useful.
Or maybe they already exist and I just haven’t seen them?