Every couple of weeks, Mark walks through the best tips and community questions from the Tana Outliner Slack. Week 28 is all about commands: three reader questions on hoisting field children, compounding set field with duplicate, and node filters, then a look at how search nodes stay in sync across every instance of a supertag.
Hoist field children into a node [0:22]
Dondry asked for a command that hoists children out of a specific field and turns them into normal nodes. The answer is a straightforward command node built on Insert Tana Paste: the prompt is a reference to the field you're moving from, and the output strategy is set to child. Run it and the field's children reappear as a plain child of the node. You can compound it further by also running set field value on a field instance with no value, which clears the original field as it moves the content out, so nothing is left behind.
Set field and duplicate, plus command line commands [3:08]
Tobias asked about another compound command. Here you set a field value, using a field instance (the > symbol) rather than a field reference, then duplicate the node. The useful trick underneath is command line commands: the built-in actions you see under command K that aren't separately configurable. You can still call them inside a custom command with "run command line command", so typing duplicate into a command chain runs the built-in duplicate. That lets you stack set field value and duplicate into one action that sets the field and produces a copy carrying the same value.
Node filters work like search nodes [5:53]
Jim's question was really about node filters, and the key insight is that a node filter works just like a search. It controls what a command is allowed to run on. With no filter, a command runs on any node; add a filter and it only appears where the conditions match. Because the syntax is identical to a query, you can draft the conditions in a search node, using the and, or, and not buttons, then drag that query into the node filter rather than copying it, which keeps it clean. On non-matching nodes the command drops out of command K, and if you place it in a supertag's commands menu it shows up as a button only when the node still satisfies the filter.
Refreshing search nodes across supertag instances [10:53]
This is an old classic that trips people up: how does Tana propagate a change to a search node that lives on a supertag template? Edit the search on a single instance and only that instance updates. To push it everywhere, run "update search node on tag and refresh all instances", which sends your change back to the template and applies it to every instance. If you're already editing the template on the tag itself, you skip the update step and run "refresh search nodes on all instances" instead. It's a bit fiddly to work out the first time, but it covers both directions cleanly.
What's coming up [15:04]
The Outliner MCP help sessions keep running on their regular cadence, and they're worth dropping into: hooking Tana Outliner up to Claude Code or Codex through MCP automates a lot of the command work covered here. The Systems Lab series builds on itself, including a recent session on loop engineering, so start with the guide section on the Outliner site if you're new to it. The big one on the horizon is remote MCP, which moves beyond the current local-only setup and lets the new Tana talk to Tana Outliner. Watch Slack and Luma for dates.