Mintplex-Labs/anything-llm
1
0
Fork 0
mirror of https://github.com/Mintplex-Labs/anything-llm.git synced 2026-07-21 01:25:28 -04:00
Code Issues 328 Packages Projects Releases 10 Wiki Activity

[GH-ISSUE #3313] [BUG]: Core Agent Skills Not Working #2132

New Issue
Closed
opened 2026-02-22 18:28:16 -05:00 by yindo · 4 comments
yindo commented 2026-02-22 18:28:16 -05:00
Owner
Copy Link
Copy Source

Originally created by @paulerbear on GitHub (Feb 21, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3313

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

SYSTEM:
Windows 11
Windows Defender Firewall turned off

Agent unable to search or web scrape. Have tested three of the web search providers (Google, DuckDuckGo, and Tavily Search)

Image
Image

Are there known steps to reproduce?

No response

Originally created by @paulerbear on GitHub (Feb 21, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3313 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? SYSTEM: Windows 11 Windows Defender Firewall turned off Agent unable to search or web scrape. Have tested three of the web search providers (Google, DuckDuckGo, and Tavily Search) ![Image](https://github.com/user-attachments/assets/60733be0-1e56-4b49-a546-47bb2cf45ba9) ![Image](https://github.com/user-attachments/assets/2cd8f171-e6fb-4ad7-9a92-e40d96702e9a) ### Are there known steps to reproduce? _No response_
yindo added the possible bug label 2026-02-22 18:28:16 -05:00
yindo closed this issue 2026-02-22 18:28:17 -05:00
yindo commented 2026-02-22 18:28:17 -05:00
Author
Owner
Copy Link
Copy Source

@Packsod commented on GitHub (Feb 22, 2025):

In theory, DuckDuckGo is online at any time without any additional conditions.
I guess if you live in Iran or somewhere like that and can only connect to major network services via VPN, then you would need to enable VPN global proxy mode to allow the search engine API to be called.

@Packsod commented on GitHub (Feb 22, 2025): In theory, DuckDuckGo is online at any time without any additional conditions. I guess if you live in Iran or somewhere like that and can only connect to major network services via VPN, then you would need to enable VPN global proxy mode to allow the search engine API to be called.
yindo commented 2026-02-22 18:28:17 -05:00
Author
Owner
Copy Link
Copy Source

@paulerbear commented on GitHub (Feb 22, 2025):

Here's an example of what attemting search with DuckDuckGo configured as @agent search looks like.
I live in South Africa - there are no major network blocks.
Image

@paulerbear commented on GitHub (Feb 22, 2025): Here's an example of what attemting search with DuckDuckGo configured as @agent search looks like. I live in South Africa - there are no major network blocks. <img width="1351" alt="Image" src="https://github.com/user-attachments/assets/4fa2281a-a90d-47ad-aaee-146c14dfce73" />
yindo commented 2026-02-22 18:28:17 -05:00
Author
Owner
Copy Link
Copy Source

@timothycarambat commented on GitHub (Feb 22, 2025):

What model are you uisng. We have written why this could happen pretty extensively across GH issues and our docs
https://docs.anythingllm.com/agent-not-using-tools

@timothycarambat commented on GitHub (Feb 22, 2025): What model are you uisng. We have written why this could happen pretty extensively across GH issues and our docs https://docs.anythingllm.com/agent-not-using-tools
yindo commented 2026-02-22 18:28:17 -05:00
Author
Owner
Copy Link
Copy Source

@paulerbear commented on GitHub (Feb 22, 2025):

I tested Deepseek R1 11B, and Llama 3.1 8B using AnythingLLM as the provider.
Perhaps you could include a list of compatible models, or flag that quantized models and context window size is an issue in the Agent Tools tooltip in app. Which model(s) tend to work well?

Also, why are you relying on prompt injection for agent skills? that doesn't seem very scalable or desirable. Why not put instructions for agent skills in a vector DB for RAG? Would that not be more robust?

@paulerbear commented on GitHub (Feb 22, 2025): I tested Deepseek R1 11B, and Llama 3.1 8B using AnythingLLM as the provider. Perhaps you could include a list of compatible models, or flag that quantized models and context window size is an issue in the Agent Tools tooltip in app. Which model(s) tend to work well? Also, why are you relying on prompt injection for agent skills? that doesn't seem very scalable or desirable. Why not put instructions for agent skills in a vector DB for RAG? Would that not be more robust?
yindo changed title from [BUG]: Core Agent Skills Not Working to [GH-ISSUE #3313] [BUG]: Core Agent Skills Not Working 2026-06-05 14:44:43 -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 possible bug
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#2132
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?