still not sold on agent orchestration. needs to be super simple and as close to the native agent interface, like Conductor et al. adding another layer of complexity while LLMs are evolving so rapidly is just fighting against the waves.

Umm... I just discovered vibe-kanban[1] as a Kanban-based coding agent and superpowers[2] for brainstorming last week, and am planning to try it this weekend. And your product looks like the combination of both... It's interesting so I hope I'll have enough energy to try all of them, but my gut tells me I'll end up spending my weekend on my bed.

[1] https://github.com/BloopAI/vibe-kanban [2] https://github.com/obra/superpowers

Premature optimization as a service? Ok that's a bit unfair but it would be reassuring if the examples showed that debugability was a design pillar.

That's fair, the PRD example is pretty cool but not the best - I have a few workflows that I use myself for doing bug bounties and deep research, but they are a bit much.

The current examples were mostly to show how you could set up different phases and how agents can work in parallel in different phases while still keeping some structure - basically any problem that you could define the logical phases in the way for solving it is going to work great with Hephaestus.

Soon I'll release a few more examples and a few tools that would make Hephaestus a development tool that learns about your repo as time goes and able to fix and add features for your code base

That's great to hear! Exciting times.

I’m curious how this would work for production level apps. My concern is that it would go off and do a lot of extra stuff and grow your codebase with orphaned code or non mvp features.

Hey, This is just a very first version of Hephaestus that is only a month in a work to showcase the idea of an autonomous agent orchestration. As we speak I'm building on top of it a development tool that would learn your codebase as time goes on, which is hopefully going to result in faster development.

Although do note that Hephaestus is not only for development purposes, it's supposed to be an agnetic framework for building workflows - such as a research workflow, investment workflow and so on, based on the semi-structured phases principles.

Also - everything is forever going to be open source (including the future features and development tools), so look out :)

Cool, nice work. I’ll check it out at some point

Very inspirational, cool! Good job!

Couldn't quite get everything to run since there was so much configuration to setup, but the idea is solid.

It's still a very first version honestly.. What did you find in the configuration hard? I'll try to streamline it so it'll be easier to run

I tried to get e2e working with Codex on a python fastAPI webapp. I made it all the way to the "organizer" who were supposed to read the PRD, but always failed since the agent started executing the PRD markdown as a python script.

My request would be to have an example PRD with premade workflow which works reasonably well for any project. I didn't use the `run_example.py` as it lacked workflows. At the end of the day, I just want to investigate if the project is something I'd like to dedicate time to, I gave up after about 2-3 hours. To have it 'plug-and-play' would be neat.

There's also a lot of references to your local filesystem, which perhaps could be cleaned up.

Terminators soon to follow or the matrix

Terminators make good project managers apparently