A place for developers to talk about building on Farcaster.

/dev

One of the things I designed and built recently I’m quite proud of, Scout Agent.

An AI telegram bot that allows you to find repos of interest in @scoutgamexyz that you can contribute to as a builder.

If you’re thinking of building an agent and would like some help, my DMs are open.

Fantasy Sports with Onchain Developers. 

Scout. Build. Win. Brought to you by the builders of SUNNYs and CharmVerse.

https://scoutgame.xyz/

One of the things I designed and built recently I’m quite proud of, Scout Agent.

An AI telegram bot that allows you to find repos of interest in @scoutgamexyz that you can contribute to as a builder.

If you’re thinking of building an agent and would like some help, my DMs are open.

https://t.me/scoutgame_developers_bot

oh wow this is so cool. im on the process of creating one for $hackathon, and the idea is for it to understand the evolution of a GitHub repo to then create a weekly leaderboard based on the progress each hacker made that week

and this frame is a frontend of that. i like the idea of agents having different frontends (TL, apps, blockchains, DCs, etc) as interfaces through which they communicate with the world

Purple 🟣 #134
Building ScoutGame & CharmVerse.io

Love to talk about new ideas, build and use AI daily

oh wow this is so cool. im on the process of creating one for $hackathon, and the idea is for it to understand the evolution of a GitHub repo to then create a weekly leaderboard based on the progress each hacker made that week

and this frame is a frontend of that. i like the idea of agents having different frontends (TL, apps, blockchains, DCs, etc) as interfaces through which they communicate with the world

https://hackathontoken.com

Yes, the nice thing about agents is that under the hood it’s all the same.

So once you have an agent on Telegram, porting it to Discord, custom web app or another interface is super easy.

End of day the only info that’s coming through is text and a user identifier.

and the agent is out there running on a vps?

how would you go about this challenge? im thinking of having the users submit the repo that they will work on, and then have a webhook that updates the agent state based on commits and what was added to the codebase

but how can the agent understand the actual functionality of the users submission?

How I would define an AI agent at the start of 2025:

In abstract, a piece of software that can achieve a certain goal, and uses AI to navigate the fuzzy path of achieving that goal.

More specifically, a combination of
- Goal
- Context management
- Underlying model
- Tools (datastores and functions with effects)
- Prompting & self-prompting strategies (including delegating to other agents)

Example: A geocoding agent

Goal: Parse content from a sentence and geocode it
Context Management: Trimmed sliding window. We only need the very most recent messages to get the job done.
Underlying model: A simple model like 4o mini is more than enough
Tools: A function to search cities with inbuilt query match ranking, and a  function with access to a Google Maps geocoding API
Prompting: Single turn. The agent should be able to get the job done in a single exchange, or ask the user to clarify the location if needed

Here’s a relevant post I made a while ago to get a comprehensive mental model on agents

Here’s a relevant post I made a while ago to get a comprehensive mental model on agents 

https://warpcast.com/meb/0x6924c7aa