Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-19 14:13:35 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #3451] [FEAT]: Shell/Script invocation block for flow builder #2229

New Issue
Closed
opened 2026-02-22 18:28:45 -05:00 by yindo · 4 comments
yindo commented 2026-02-22 18:28:45 -05:00
Owner
Copy Link
Copy Source

Originally created by @yonil7 on GitHub (Mar 13, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3451

What would you like to see?

Custom agent skills documentation say: "you can create custom agent skills that can run extra processes like running a local Python script ..."

Some suggestions:

  1. The documentation can include an example custom agent skill handler.js file that shows how to call a python script from js
  2. anything-llm can be enhanced to allow integrating and running any executable as a custom agent skill by composing plugin.json and setting the entrypoint in it to the executable to invoke. (e.g. python my_script.py or bash -c my_command, ...)

My specific use case for this (generic) requirement: I would like to be able to use browser-use python package for automating/executing tasks on my currently open browser tabs (e.g. extract a table from a site after I've already logged in manually)

Originally created by @yonil7 on GitHub (Mar 13, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3451 ### What would you like to see? Custom agent skills documentation say: "you can create custom agent skills that can run extra processes like running a local Python script ..." Some suggestions: 1. The documentation can include an example custom agent skill `handler.js` file that shows how to call a python script from `js` 2. `anything-llm` can be enhanced to allow integrating and running **any** executable as a custom agent skill by composing `plugin.json` and setting the `entrypoint` in it to the executable to invoke. (e.g. `python my_script.py` or `bash -c my_command`, ...) My specific use case for this (generic) requirement: I would like to be able to use [browser-use](https://github.com/browser-use/browser-use) python package for automating/executing tasks on my currently open browser tabs (e.g. extract a table from a site after I've already logged in manually)
yindo added the enhancementfeature request labels 2026-02-22 18:28:45 -05:00
yindo closed this issue 2026-02-22 18:28:45 -05:00
yindo commented 2026-02-22 18:28:45 -05:00
Author
Owner
Copy Link
Copy Source

@shatfield4 commented on GitHub (Mar 13, 2025):

Unfortunately we will not be supporting python for this.

@shatfield4 commented on GitHub (Mar 13, 2025): Unfortunately we will not be supporting python for this.
yindo commented 2026-02-22 18:28:45 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 13, 2025):

To elaborate, this kind of skill would be done via the "script execution" block in the flow builder - which is what you are describing. Where we use the systems python install and then invoke that to run a script you also have built.

This is actually being done already - we can use this issue to track that items progress so at least there is a public endpoint now. If you mean writing python code inside anythingLLM for a custom agent skill, well that is quite unlikely since we are a NodeJS project so we cannot execute python scripts.

If you want to use python in a custom skill where you write code directly, the only thing you can do is write you python script then you can use spawn in the NodeJS handler.js and invoke the python script using the systems python executable. We wont ever pre-bundle python into AnythingLLM though automatically

@timothycarambat commented on GitHub (Mar 13, 2025): To elaborate, this kind of skill would be done via the "script execution" block in the flow builder - which is what you are describing. Where we use the systems python install and then invoke that to run a script you also have built. This is actually being done already - we can use this issue to track that items progress so at least there is a public endpoint now. If you mean writing python code _inside_ anythingLLM for a custom agent skill, well that is quite unlikely since we are a NodeJS project so we cannot execute python scripts. If you want to use python in a _custom skill_ where you write code directly, the only thing you can do is write you python script then you can use `spawn` in the NodeJS handler.js and invoke the python script using the systems python executable. We wont ever pre-bundle python into AnythingLLM though automatically
yindo commented 2026-02-22 18:28:45 -05:00
Author
Owner
Copy Link
Copy Source

@yonil7 commented on GitHub (Mar 14, 2025):

Thanks for the elaboration.
My suggestion is about the ability to run any executable/process (written in any language) as a custom agent skills (not about being able to run python code directly inside anythingLLM - I also don't think this make sense)

The suggestion is mainly about enhancing the documentation on this existing functionality:

  • For no-code users (using agent flows blocks) - add "script execution" block to the documentation (its not there)
  • For developers building custom agent skills - the documentation can include an example handler.js that just shows how to run any other executable/process (e.g. spawn) (for non js developers)

(in the future, plugin.json:entrypoint semantic/schema can be enhanced to allow the user to set any executable/process to run (e.g. "bash -c my_command") instead of only js files)

@yonil7 commented on GitHub (Mar 14, 2025): Thanks for the elaboration. My suggestion is about the ability to run **any** executable/process (written in **any** language) as a custom agent skills (not about being able to run python code directly inside anythingLLM - I also don't think this make sense) The suggestion is mainly about enhancing the documentation on this **existing** functionality: * For no-code users (using agent flows blocks) - add "script execution" block to the documentation (its not there) * For developers building custom agent skills - the documentation can include an example `handler.js` that just shows how to run any other executable/process (e.g. spawn) (for non `js` developers) (in the future, `plugin.json:entrypoint` semantic/schema can be enhanced to allow the user to set any executable/process to run (e.g. `"bash -c my_command"`) instead of only `js` files)
yindo commented 2026-02-22 18:28:45 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Mar 17, 2025):

Then we can track this in https://github.com/Mintplex-Labs/anything-llm/issues/3484

@timothycarambat commented on GitHub (Mar 17, 2025): Then we can track this in https://github.com/Mintplex-Labs/anything-llm/issues/3484
yindo changed title from [FEAT]: Shell/Script invocation block for flow builder to [GH-ISSUE #3451] [FEAT]: Shell/Script invocation block for flow builder 2026-06-05 14:45:17 -04:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Desktop
Docker
Integration Request
Integration Request
OS: Linux
OS: Mobile
OS: Windows
UI/UX
blocked
bug
bug
core-team-only
documentation
duplicate
embed-widget
enhancement
feature request
github_actions
good first issue
investigating
needs info / can't replicate
possible bug
pull-request
Mirrored from GitHub Pull Request
 question
stage: specifications
wontfix
No Label enhancement feature request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#2229
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?