Feature Requests

Standard agent specialised in generating template-driven technical and user documentation (technical writer quality) of the solution.

Agents specifically primed to be trained on, and have context access to, existing git repos (instead of focusing on creating from scratch), and amending or updating that.

Sync generated tasks and sub-tasks back to Jira

When I try to add a notebook it tells me to add a project first. You removed that option in the latest 1.2.0 update, and have replaced it with a listing for Repository that does nothing when clicked (on my Win10 machine). You are releasing an app that's basically a non-functional menu interface from what I can tell. Did anyone even open this up and check to make sure it was functional on Win10 before releasing it?

Hmm, I was wondering if anyone here knows - is there a way to develop an agent that manages other agents by delegating tasks to them as needed? Also, how do Fine think the management of agents may progress going forward? Do you think there's potential for agents to self-report and self-grade their work, allowing them to autonomously improve their performance over time without human oversight? Thanks 🙏

Ability to run agents locally (own machine or cloud hosting) to keep material private or just to use own infrastructure.

An agent type that generates a POC UI or wireframe or even just architecture diagrams, in real-time, showing its understanding of what is described in the tech spec while the user is still fleshing it out.

Transform visual prototypes (from Figma) into good frontend code (React).

Hi everyone, I have been thinking about a nice feature for Fine: a crowdsourced knowledge base of useful YAML manifest sections for agent configuration. Am I correct in my understanding that the manifest structure is not well-defined because agent functionalities are not well-defined?