Agent Kanban has 4 main features:

GitOps & team friendly kanban board integration inside VS Code Structured plan / todo / implement via @kanban commands Leverages your existing agent harness rather than trying to bundle a built in one .md task format provides a permanent (editable) source of truth including considerations, decisions and actions, that is resistant to context rot

TLDR: sql.js loads SQLite into a WASM module with a single linear memory buffer. After indexing 18,000+ files that memory is large and fragmented. Reconstructing a Database() object on that same heap fails with "memory access out of bounds". The fix is a build-time esbuild plugin that injects a resetCache() function into the sql.js bundle, allowing the WASM module to be fully discarded and reloaded from scratch on rebuild.

Even with the best models, agent-assisted coding can be hit and miss. A good AGENTS.md that covers most of the mistakes you find agents making can go a long way, but this flow has been producing outsized results over the last couple of weeks.

It takes a little more discipline than hacking away in the chat window until you get what you want, but this process makes a "one shot" success more likely, and really pays off when things get complicated and the agent needs more of a steer.

AS Notes brings Obsidian/Logseq-style knowledge management directly into VS Code - wikilinks, backlinks, page aliases, daily journal, and a task panel - without leaving your editor or touching a separate app.

The motivation was simple: most developers already take notes, but the friction of switching to another app means those notes slowly drift into neglect. AS Notes keeps everything in plain .md files + wikilinks, in a folder you control. There's no cloud sync, no accounts, no telemetry. The index is a local SQLite database powered by WASM (no native dependencies), so it works everywhere VS Code does. Because your notes are just markdown files, they get the same Git workflow as your code.

The feature I'm most pleased with is nested wikilink handling - `[[Project [[Alpha]] Notes]]` creates multiple navigable targets from a single link, which mirrors how Logseq-style linking naturally composes. There's also rename tracking: rename a file and AS Notes offers to update every reference across your workspace.

Install from the VS Code Marketplace or clone the demo notes repo to kick the tyres. Happy to answer questions about the architecture - particularly the WASM SQLite indexing approach.

GitHub: https://github.com/appsoftwareltd/as-notes

Marketplace: https://marketplace.visualstudio.com/items?itemName=appsoftw...

Comments URL: https://news.ycombinator.com/item?id=47135694

Points: 1

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=47113708

Points: 10

# Comments: 1

My Lenovo ThinkPad P1 Gen 4 is coming up for 4 years old. It is a powerful workstation, and has a good, but by no means state of the art GPU in the RTX A3000. My expectation is that many developers will have a PC capable of running local LLMs as I have set up here.

See the GitHub repository for the full walk through:

https://github.com/gbro3n/local-ai/blob/main/docs/local-llm-...

Ref: https://www.appsoftware.com/blog/local-llm-setup-on-windows-...

Comments URL: https://news.ycombinator.com/item?id=47110823

Points: 4

# Comments: 3

I've also noted such a huge gulf between some developers describing 'prompting things into existence' and the approach described in this article. Both types seem to report success, though my experience is that the latter seems more realistic, and much more likely to produce robust code that's likely to be maintainable for long term or project critical goals.