[GH-ISSUE #5356] [NOBUG - CLARIFICATION] - @agent DOA via api in 1.12.0?? #5041

Closed
opened 2026-06-05 14:51:40 -04:00 by yindo · 1 comment
Owner

Originally created by @ohromeo on GitHub (Apr 5, 2026).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5356

Hi,

not sure but could swear in previous docker version i was able to access @agents via my front-end using the API's.
now with 1.12.0 the API wants nothing to do with them, my default prompts return its response.
my slash commands are working. checking the network calls when sending the @agent requests appear to execute and return 200:

when i look at the system prompt in the anythingLLM desktop (docker/linux) - i see my sytem prompt as i have set it via the api, and when executing the same @agent via the desktop.. it works but not through the API calls

not sure if i'm just dreaming the @agents worked via my API calls, or if something is broken or my system prompt is too restrictive or incorrect?

EDIT: i think this #1484 in docs noted that it's not supported --- not sure this is still applicable..

thanks
Johnny

Originally created by @ohromeo on GitHub (Apr 5, 2026). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/5356 Hi, not sure but could swear in previous docker version i was able to access @agents via my front-end using the API's. now with 1.12.0 the API wants nothing to do with them, my default prompts return its response. my slash commands are working. checking the network calls when sending the @agent requests appear to execute and return 200: when i look at the system prompt in the anythingLLM desktop (docker/linux) - i see my sytem prompt as i have set it via the api, and when executing the same @agent via the desktop.. it works but not through the API calls not sure if i'm just dreaming the @agents worked via my API calls, or if something is broken or my system prompt is too restrictive or incorrect? EDIT: i think this #1484 in docs noted that it's not supported --- not sure this is still applicable.. thanks Johnny
yindo closed this issue 2026-06-05 14:51:40 -04:00
Author
Owner

@ohromeo commented on GitHub (Apr 5, 2026):

i believe the issue has been resolved. looks like my code sequence for @agent commands wsa incorrect:
"Please review this policy... \n\n @agent list your tools?"

changed my code as:

Without @agent: "Please review this policy... \n\n What's our leave policy?"

With @agent: "@policy-expert What's our leave policy? \n\n Please review this policy..."

The agent sees its command first, processes it correctly, AND still gets the user_context for reference.

<!-- gh-comment-id:4189843817 --> @ohromeo commented on GitHub (Apr 5, 2026): i believe the issue has been resolved. looks like my code sequence for @agent commands wsa incorrect: "Please review this policy... \n\n @agent list your tools?" changed my code as: Without @agent: "Please review this policy... \n\n What's our leave policy?" ✅ With @agent: "@policy-expert What's our leave policy? \n\n Please review this policy..." ✅ The agent sees its command first, processes it correctly, AND still gets the user_context for reference.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#5041