langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-21 03:55:25 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

Considering a 3rd mode: Agentkit agent #1258

New Issue
Closed
opened 2026-02-21 17:31:32 -05:00 by yindo · 0 comments
yindo commented 2026-02-21 17:31:32 -05:00
Owner
Copy Link
Copy Source

Originally created by @pascalwhoop on GitHub (Mar 7, 2024).

Self Checks

  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
  • Pleas do not modify this template :) and fill in all the required fields.

1. Is this request related to a challenge you're experiencing?

The current agent framework used is fairly difficult to customise. There are many situations where a business application needs to follow fairly rigid process steps. For example, a KYC process at a bank needs to collect information from 3 to 4 specific sources before aggregating them together and evaluating the content with an LLM. This sort of sequence of steps is very difficult to program with Dify

2. Describe the feature you'd like to see

There are a few projects out there which allow a more structured definition of what steps the agent should take. One example is
https://github.com/BCG-X-Official/agentkit
This would allow a more controllable behaviour of the agent which generally leads to higher quality outputs. Maybe this framework is also a not ideal example. My point is there should be a form of DAG style set of actions that could be defined and chained so that the agent performs a set of predefined steps. E.g.:

  1. retrieve user data from a specific DB
  2. start researching about the person online to find information
  • search in factiva
  • search in google
  • search in proprietary API X
  1. translate all answers to english for each non english result
  2. consolidate all content
  3. write a report
  4. store report by calling a specific API

3. How will this feature improve your workflow or experience?

It would allow for better instructability of the agent.

4. Additional context or comments

No response

5. Can you help us with this feature?

  • I am interested in contributing to this feature.
Originally created by @pascalwhoop on GitHub (Mar 7, 2024). ### Self Checks - [X] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [X] I confirm that I am using English to submit this report (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [X] Pleas do not modify this template :) and fill in all the required fields. ### 1. Is this request related to a challenge you're experiencing? The current agent framework used is fairly difficult to customise. There are many situations where a business application needs to follow fairly rigid process steps. For example, a KYC process at a bank needs to collect information from 3 to 4 specific sources before aggregating them together and evaluating the content with an LLM. This sort of sequence of steps is very difficult to program with Dify ### 2. Describe the feature you'd like to see There are a few projects out there which allow a more structured definition of what steps the agent should take. One example is https://github.com/BCG-X-Official/agentkit This would allow a more controllable behaviour of the agent which generally leads to higher quality outputs. Maybe this framework is also a not ideal example. My point is there should be a form of DAG style set of actions that could be defined and chained so that the agent performs a set of predefined steps. E.g.: 1. retrieve user data from a specific DB 2. start researching about the person online to find information - search in factiva - search in google - search in proprietary API X 3. translate all answers to english for each non english result 4. consolidate all content 5. write a report 6. store report by calling a specific API ### 3. How will this feature improve your workflow or experience? It would allow for better instructability of the agent. ### 4. Additional context or comments _No response_ ### 5. Can you help us with this feature? - [X] I am interested in contributing to this feature.
yindo added the 💪 enhancement label 2026-02-21 17:31:32 -05:00
yindo closed this issue 2026-02-21 17:31:32 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label 💪 enhancement
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: langgenius/dify#1258
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?