← Back to the Marketplace

Daily picks · November 19, 2026

Bots That Tidy Up Your Digital Life

From a warm butler persona to an Obsidian librarian, a multi-bot dispatcher, and a bug-hunting partner, today's picks all specialize in the quiet work that keeps everything else running.

A warm, ready-made assistant persona

A complete personality and appearance package for Tobs that aims to feel like a warm, capable butler rather than a chatbot. The persona runs on seventeen sections covering everything from speaking rhythm to memory boundaries, with three hard rules: never fabricate, never bluff, never perform. It is the top trending listing this week, and the packaged avatar, banner, and color scheme make it a one-stop install for anyone who wants a ready-made assistant persona.

大肥鱼

A dispatcher for multi-bot projects

If you coordinate several bots, this one positions itself as the general manager above them: it restates the final deliverable, asks every open question once, splits work into single-responsibility tasks with checkable completion criteria, dispatches each job to the cheapest or most capable bot, and verifies results against evidence before reporting back. The listing frames it as the backbone of a delivery pipeline, which makes it a notable pick in this week's hot section.

言

An Obsidian vault that keeps itself tidy

For Obsidian users drowning in notes, this bot applies the LLM-Wiki approach to keep a vault tidy: it preserves original sources and writes summaries, answers questions with cited evidence and disclosed uncertainty, and runs inspections for broken links, orphan pages, duplicates, and conflicts. It promises to back up before changing anything and never silently overwrite conflicting sources, a reassuring guardrail for a tool with write access.

Obsidian 知识库治理

A bug partner that reproduces before fixing

New to the marketplace, Debugging Partner is built for the moment a bug report is too vague to act on. It insists on writing the failure as a testable claim, gets a reproduction before proposing anything, prefers one experiment that separates two hypotheses over a pile of hopeful edits, and verifies both the original failure and a regression check before calling it fixed. If you want a coding companion that slows down to get it right, this is one to watch.

Debugging Partner

Past picks