This looks very similar to what I've seen with a different harness - https://github.com/AMAP-ML/LongHorizon-Harness. It's critique pattern looks like another version of lh-harness without looping toward an overarching goal
It does matter though. If you want to murder someone by hitting them with a plushie, you're not going to get charged with attempted murder because it's not possible that that would ever work. There must be justification that the choice will have the intended effect.
We should not be gung-ho to give the government more power to regulate speech.
It isn't so clear there is a reason for the crime of 'attempt', or what or how good the reason is. The Star Chamber, which cooked it up, is universally condemned as oppressive. You might say they anticipated science fiction tyranny, when they invented the world's first thought-crime: they were able to make a crime of 'conspiracy' because they imagined that the speech of the conspirators was an 'outward act'; inevitably its glorious future was e.g. to jail the left wingers in the McCarthy period.
The point of view that says 'there is a reason we punish attempts' has difficulty explaining why we punish 'success' _even more_. There is a sort of bad concience about it. The paradoxes are discussed in a characteristically brilliant and twisted work 'The Punishment that Leaves Something to Chance' by David Lewis, one of the greatest philosophers of the 20th c. It nominally defends the law of attempt but can as well be read as a catastophically destructive parody of it. https://andrewmbailey.com/dkl/Punishment_Chance.pdf
It's interesting to claim that it would be difficult to explain why we punish successful crime more. A successful murder creates more suffering (victim's friends and family), and ostensibly the loss of a productive member of society. These all seem like fairly common justifications for punishing murderers under the banner of retributive justice.
There's even a community plugin MCP that makes this easier. Afaik the hot reload is the only thing that "beats" this but I haven't had issues with Agents interacting with obsidian through the MCP - I'm not sure how much more useful it is to talk with an agent through a note taking app vs through its CLI.
Obsidian does hot reloading too. Usually, I have vscode/Claude Cowork up on one side and on the other is Obsidian. I made a plugin[1] that converts emails to `.md` so they can also be in Obsidian folder -- so Claude can query those emails the same way it does my notes. Works beautifully.
