I use it for all my personal projects. Any new project where I will have todo lists, I'll run `kan init` and `kan serve`, and begin tracking tasks. Because it's a CLI, you can interact with the board thru the CLI with commands like `kan add` and `kan edit`, so I've even included a SKILL markdown file in the repo which I use to let Claude read and interact with my boards. You can also navigate all your boards from a single `kan serve`, so it's convenient and snappy cross-project too.
Ask HN: What are you working on? (February 2026)
941–950 of 1001 posts
Re: Ask HN: What are you working on? (February 2026)
#942So I built Tasty A.F. — paste a recipe URL, and it uses AI to extract the essentials and format them onto a clean, printable notecard.
How it works:
1. Paste any recipe URL
2. AI scrapes the recipe and generates a concise 3x5 card
3. Print it, stick it on the fridge or a notecard, and cook
Built with Python/Flask and the Anthropic Claude API, hosted on Railway.
Re: Ask HN: What are you working on? (February 2026)
#943Re: Ask HN: What are you working on? (February 2026)
#944Re: Ask HN: What are you working on? (February 2026)
#945- https://github.com/desplega-ai/qa-use CLI to control browsers and tests locally, coding agent friendly - https://github.com/desplega-ai/agent-swarm OSS agent swarm that can be deployed with docker compose using Claude - https://desplega.ai platform that supports qa-use backend
Cheers!
Re: Ask HN: What are you working on? (February 2026)
#946Re: Ask HN: What are you working on? (February 2026)
#947https://github.com/WispFramework/Wisp, https://wispframework.github.io/Wisp/
It tries to be reasonably lightweight but batteries-included, with extra features in separate "extensions" packages. It's also designed to be modular. All functionality has a default implementation but can easily be replaced if desired. It uses the Fluid template engine[1] (a .NET implementation of the Liquid language).
It's been a great way to dig into more advanced concepts like reflection and HTTP internals and while it's probably not safe to use in production, I have used it to build small private apps with great success.
The current major pain points are the lack of websocket support, which I'm planning to add soon, and the general fragility of the NetCoreServer[2] based HTTP backend, which I'm planning to replace with EmbedIO. (Ideally, I would love to use Kestrel here, but it's so deeply baked into ASP.NET Core that you can't use it without pulling in the whole Web SDK)
The documentation is also pretty incomplete but I hate writing docs so I find it hard to force myself to do it.
[1] https://github.com/sebastienros/fluid
[2] https://github.com/chronoxor/NetCoreServer
(LLM Disclaimer: This project is not vibe-coded. Most of the code is written by hand, with some input from ChatGPT being used as a fancy search engine. The docs are written by hand and then fed through ChatGPT to make them more readable)
Re: Ask HN: What are you working on? (February 2026)
#948By embedding realistic decoy routes and honey fields that are difficult to distinguish from real API constructs, attackers are nudged to authenticate — converting reconnaissance into actionable security telemetry.
github: https://github.com/trappsec-dev/trappsec
docs: https://trappsec.dev
Re: Ask HN: What are you working on? (February 2026)
#949The agent uses LLMs + RAG to understand product descriptions and classify them hierarchically into 6, 8, or 10-digit HS/HTS codes, including country-specific variants.
Key points: – We built this after seeing how much time and money teams lose due to misclassification and customs delays.
Would love feedback on: • Accuracy expectations • Edge cases / ambiguous products • How this compares to existing customs workflows
Happy to answer questions
https://simplai.ai/agents-library/hs-code-classification-age...