fucory

1/ Smithers is the best Orchestrator and you should try it What is Smithers Smithers is an Orchestration framework where you declaratively create a DAG of Durable Tasks as xml and reactively regenerate parts of that DAG as Tasks complete How do you use it You use it as a CLI or a agent skill. Your agent then writes, updates, and monitors those scripts What special features does it have - MDX for prompts - Library of prebuilt 0-config workflows - All workflows represented as code that is easy to modify, customize, or compose into larger workflows - Built in observability with SQLLite and pormetheus - Supports Claude code, Codex, Gemini, and Kimi subscriptions - Supports all harnesses including Claude Code, Codex, Pi, and Opencode - Supports Agent SDK Why is it so good? The main reasons Smithers is so good is 1. Agents are great at writing complex workflows because Agent's are great at React 2. It is thoughtfully built with @EffectTS_ Durable execution primitives enabling more reliability for very long running agentic tasks 3. Powerful CLI for watching and interacting with wrokflows 4. MDX is a really great existing standard for expressing prompts 5. The declarative nature of Smithers and it's great observability makes it very easy to read and quickly reason about What is best way to get started? bunx smithers init in an existing project and try it out Alternatively check out the docs smithers.sh

We are looking to help integrate smithers into your project! If you are using smithers 1. Add your project to awesome smithers: github.com/smithersai/awe… 2. Reach out if you want me to personally help get your company integrated with Smithers

Smithers now has a preferred sandbox provider @microsandbox. It's fully open source allowing us in future releases to easily build smithers specific integrations

Forget loops, forget graphs State Machines + memory Smithers 0.29.0 is out. 225 commits, 726 files changed, 84,604 insertions. 🧠First-class agent memory, ♻️ XState state machine support 📦Fully OSS MicroVMs.

@AlecFreese @shawmakesmagic Very dank

@FUCORY @shawmakesmagic the pathway from slop to dank is design. Obsessive craftsmanship and intentionality around what you build

The opposite of slop is dank. One can magically turn slop into dank via caring a lot about how dank it is

@eniwhere_ Pass the skills in as input strictly better

I actually think self invoking skills are a bad thing

@dexhorthy I've been turning this general way of working into a workflow+UI for the past 4 weeks. It's such a great way of working

today's fun ai coding tricking - when doing the high-level part of your plan, do an amazon-style "working backwards" approach - write the customer-facing blog post about the feature BEFORE you start building this can surface so many edge cases and help you dial in on prioritizing the important parts, and just is much nicer to read (plus you can post it after you ship 🙂)

@_Felipe Yea Smithers.sh is built in typescript but long term I want to move to supporting any language but recommending starlark as the default

@FUCORY Yeah. You’ve read my mind.

We need a programming language.

@_Felipe Like starlark?

@FUCORY Too unrestricted. We need a language that can command workers and LLMs without security holes.

@gakonst @0xfoobar In past vibes was the best eval because things moved fast and we were nowhere close to optimal. Now evals are everything

imo the biggest issue in the ai tooling ecosystem (largely pushed by the token merchants) is the relentless promotion of SKILLs, subagents, loops and all sorts of "vibes" techniques without accompanying evals

@avrldotdev You can actually decrease cost via workflows if decreasing cost without making time and quality worse is a goal

i never used loops, workflows, ultracode, subagents, agent swarms, all that hype mumbo jumbo. I work in a chat window with all the tools available & it's more than enough. For 95%+ of devs, it's all just a waste of tokens & no extra benefits, made to increase your cost!

@broady Smithers.sh uses flux or you can use xstate

tell me when the people discover state machines

@odysseus0z Always has been

LangGraph has always been right???

@Vtrivedy10 This is kinda how smithers.sh works

agents executing dynamically generated langgraph 👀

@Hiteshdotcom We shipped graph engineering in January. Now in July it's a mature framework. The next step after graph engineering which smithers has almost reached is you just talking to the agent while the agent engineers the graph for you

Graph engineering is here 😂😂😂

@mattzcarey Exactly! We built dynamic workflows in January with smithers.sh and users report it being the biggest unlock to their agentic journey

dynamic workflows are graphs on demand let the agent build a control flow.

@sydneyrunkle @LangChain Langchain is the og 🙏

at @LangChain we power our loops w graphs! deepagents is built on langgraph!