vistdev
Re:Yes, and
on Monday October 12, 2026 @05:49AM (#50053894)
That seems to work best (plan with the expensive model, cheap ones implement). The "Implemented feature XX, tests pass" summary drove me nuts too, so I make the planner record the decision and the why in my second brain app before anything gets built, and implementers report back against it.
Re:Show HN: Offrun – manage every coding agent from one workspace
on Tuesday October 06, 2026 @04:30AM (#49975772)
Not really an orchestrator, but my second brain notes & tasks app Vist (https://usevist.dev) has a memories system built in. You connect it via MCP, and define the projects you want to keep track of yourself. Your agents automatically update the memory/project state when needed, and this works nicely for me in a similar setup with separate repos for the Vist app, the Desktop apps, public releases, and so on.
Re:Agents don't need memory. They need documentation
on Monday October 05, 2026 @11:46AM (#49966404)
I like your approach of keeping things a simple and light-weight as possible, and asking Claude to set up its own environment works remarkably well. It even does a great job of configuring OpenCode if you ask it to ;-)
That said - I personally ended up saving my plans and status in my second brain app via MCP. Having that available to me on the road when I'm in the car brainstorming with Claude over voice is a big win.
Re:Dots: Always-on agents
on Monday October 05, 2026 @10:35AM (#49965455)
I ran into similar issues when I was still using projects (stopped that almost entirely).
I ended up pulling the cross-project stuff (the todo list, guidelines, things about me) out of project memory entirely and into my MCP-its-just-markdown-notes second brain.
Adding an instruction to my Claude.md and general instructions to call the load_context tool from my MCP server before doing anything else does sort out memory issues quite thoroughly, with very limited token creep.
Re:Jevmem – automatic project memory for Claude Code, built on Jev
on Monday September 28, 2026 @07:16AM (#49876285)
I feel that - the contents are the useful part (decisions, open questions, what’s next) and I got tired of them living in a harness-specific memory store I can’t read or edit without jumping through hoops.
One of the first patterns to emerge here is still the best for me personally: one Markdown notes app both the human and whichever agent I’m in write to. Curious what ‘where else’ looks like for you — repo file, Obsidian, or something hosted?