Alexander Wolf
@AWolf81
Followers
1K
Following
2K
Media
75
Statuses
886
Building an AI-powered VS Code extension π¨βπ» | Creator of https://t.co/JJFLayy2Es (on hold) | Twitter Banners @Gumroad | FindYourNextFavorite YT #Tech #Dev #BuildInPublic
Stuttgart, Deutschland
Joined September 2014
Created a @claudeai skill based on the Universe of Thoughts framework / paper. You can get the skill in the link below. It changes reasoning of Claude because it follows a different reasoning schema. IMO great for new buisness ideas. #dev #AI
https://t.co/6JC3MQqtnm
gist.github.com
Universe Of Thoughts Skill. GitHub Gist: instantly share code, notes, and snippets.
0
0
2
Sure, you can also use the automation in the N8N cloud but I wanted to try self-hosting n8n for personal experiments. If you'd like to know more about the setup, I can add the Docker-Compose.yml to a Github Gist. πThanks @ Maxim Osipovs https://t.co/1nSFK6Sspm
n8n.io
This n8n workflow template implements an intelligent research paper monitoring system that automatically tracks new publications in ArXiv's Artificial...
0
0
0
Self-hosted n8n at @hostinger with Dockge (VPS, KVM 1 for 5$/Month) in Docker container. It is a scraping automation that generates daily summaries of AI papers with Gemini AI and sends them to your inbox. π Link
2
0
3
My "work in progress" is on a feature branch - latest version from screenrecording not pushed yet. The simple-agent CLI is a learning experiment and an usage example for simple-rpc-ai-backend. Still a lot to test - especially skill sandboxing. https://t.co/o2c0kijM8i
0
0
0
Progress update for simple-agent. My user-interaction skill is almost working - just need to fix XML parsing, after my usage limit break...π Learned a lot about agent issues during this (confirmation loops, missing context, execution before confirmation - just to mention some)
1
0
2
Ah, OK, the files in /tmp folder are likely generated by bash tool execution for CWD tracking. Unfortunately, there is no source code in the Github repo to it. But there is a GH issue with some details about it: https://t.co/9goXTSL0uv
github.com
Bug Description Claude Code creates temporary files to track working directory changes across Bash command executions but never deletes them, causing accumulation of /tmp/claude-*-cwd files. Enviro...
0
0
1
I'm using the project root in my file-handling skill, so ./ is referencing to the folder where the simple-agent was started. Maybe I have to rethink path handling & move it to the caller. https://t.co/jrauc5jQKP
github.com
Platform-agnostic RPC server for AI integration in VS Code extensions, web apps, and CLI tools. - AWolf81/simple-rpc-ai-backend
0
0
0
It works but probably not the best way - "pollutes" the environment. But that's what Claude Code implemented.
1
0
0
The problem process.cwd() is not giving you the root of the current project in all cases. So I like the file approach and I need to recheck my current implementation. I'm creating an env. variable PROJECT_ROOT on CLI / server start, so it's accessible everywhere it is needed.
1
0
0
Interesting to see that Claude Code CLI had similar issues with the project root handling, just look at your /tmp folder there are many 'claude-<githash>-cwd' files - content is the working dir. Likely created everytime you open Claude Code extension or CLI. #buildinpublic π
2
0
2
Thanks for sharing/creating. I've also seen it in your newsletter and I'll give the :exacto models a try. Still working on my agent implementation. Just tried the qwen/qwen3-coder:exacto model. π Exactly what I need right now - more models to try and not just Claude Sonnet.π¦Ύ
Introducing `:exacto`, Precision Tool-Calling Endpoints OpenRouter now offers a curated subset of providers on top open source models with measurably higher tool calling accuracy delivering more reliable tool use.
0
0
1
What skill are you planning to create? Or have you already create one? Would love to see your ideas.π§ I think I'll start with file access skill, I already have the base but need to do more testing.
0
0
0
Maybe I'll revisit that topic later or just use @openai/codex-sdk as it is Apache-2.0 licensed. Right now, I'm sticking to my current plan - no specific provider SDK. Sure starting with Anthropic Sonnet via Openrouter as I can compare skill usage/implementation with Claude Code.
1
0
0
I understand why it is there, they want to protect their IP. I would have used it with an Apache or MIT license. But OK, then I have to put some more work into it without the SDK. Just using Vercel AI SDK. So be careful if you're planning to use it, to avoid legal issues.
1
0
0
Why I'm not using the Anthropic Claude Code SDK for my library or Simple-Agent? I have license concerns. π± There is no specific license, just a copyright + terms. And in the terms under D4, there is an exclusion for competing agents. https://t.co/UkYHoG9wvQ
#buildinpublic #dev
anthropic.com
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
1
0
3
Working on agent feature for simple-rpc-ai-backend right now. There will be also a basic command line agent called 'simple-agent' - not a full-blown CLI yet but for testing agents. It will use skills with sandboxing. Sandboxes not tested yet. #buildinpublic #devlopment #ai
0
0
2
Claude can now do work for you in your exact style and using your exact process. You can "teach" Claude to run a pre-designated workflow on any task - making agents MUCH more useful (time saved on manual formatting alone is huge). Enter...Skills π
Today we're introducing Skills in claude dot ai, Claude Code, and the API. Skills let you package specialized knowledge into reusable capabilities that Claude loads on demand as agents tackle more complex tasks. Here's how they work and why they matter for the future of agents:
23
58
851
If you'd like to give it a try, please refer to the quickstart guide below. If anything is unclear or not working, please let me know here on Twitter or add an issue at Github. π https://t.co/wLsslEwgDY
awolf81.github.io
Developer documentation for the Simple RPC AI Backend server.
0
0
0