Skip to content

Bug / UX Issue: Delegate Dialog Is Missing Critical Configuration Options #116

Description

@malys

Summary

The Kanban Delegate to Agent dialog is missing important parameters and is currently difficult to use effectively.

The dialog exposes fields such as:

  • Agent
  • Model
  • Effort
  • Mode

but key information is missing and the workflow is incomplete.

Screenshot

Attached.

Current Issues

1. Agent List Is Empty

The dialog displays:

No agents configured (add via Settings)

but does not explain:

  • Which agents are required
  • How to create them
  • Why built-in providers cannot be selected directly

This makes delegation difficult to discover.


2. Model Selection Is Unclear

The Model field is present but does not clearly expose:

Claude
Codex
Gemini
OpenAI
...

or provider-specific models.

It is not obvious:

  • Where the models come from
  • Which provider will execute the task
  • Whether the selected model belongs to Claude, Codex, Gemini, etc.

3. Effort Selection Lacks Context

The Effort dropdown is displayed but provides no guidance about:

Low
Medium
High

or cost/performance impact.

Users cannot easily determine which level is appropriate.


4. Missing Provider Selection

A dedicated field would be useful:

Provider:
[ Claude ▼ ]

or

Provider:
[ Codex ▼ ]

before selecting a model.

Example:

Provider
  └─ Claude
      ├─ Sonnet
      └─ Opus

  └─ Codex
      ├─ Sol
      └─ Terra

5. Impossible to See Effective Configuration

Before launching the delegation, users should be able to verify:

Provider: Codex
Model: Terra
Effort: High
Mode: Sync

Currently the final configuration is not obvious.


6. No Visibility After Delegation Starts

Once delegation is launched:

  • No dedicated terminal appears.
  • No live logs are shown.
  • No easy way to monitor execution.
  • No clear link between the delegated session and the parent task.

Users cannot easily understand what the delegated agent is doing.


Expected Behavior

The dialog should clearly expose:

Provider
Model
Effort
Agent
Mode

with dynamic values loaded from available providers.

The user should know exactly:

  • Who will execute the task
  • Which model is used
  • Which reasoning effort is applied
  • How to track progress afterwards

Suggested UI

Delegate Task

Provider:
[ Codex ▼ ]

Model:
[ Terra ▼ ]

Effort:
[ High ▼ ]

Mode:
[ Sync ▼ ]

Task:
[________________]

Benefits

  • Easier delegation workflow.
  • Better discoverability.
  • Better support for multiple providers.
  • Reduced configuration errors.
  • Clearer execution tracking.

Priority

High

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions