Skip to content

feat: per-task-type model assignment with three-tier system #185

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
CaitlynByrne wants to merge 2 commits into
base: master
Choose a base branch
from
Open

feat: per-task-type model assignment with three-tier system #185

CaitlynByrne wants to merge 2 commits into from

Conversation

@CaitlynByrne
Copy link

@CaitlynByrne CaitlynByrne commented Feb 9, 2026

Summary

  • Adds per-role model assignment allowing different models for each task type (initialization, coding, testing, spec creation, expansion, assistant, log review)
  • Implements a three-tier system (high/mid/low) with provider-specific defaults so roles get sensible models out of the box
  • Global model selector now sets all role assignments at once, shows a toast confirmation, and auto-expands the per-task section for further customization
  • Global model button only highlights when all roles are set to the same model, avoiding misleading state

Test plan

  • Open Settings, verify no model is highlighted in the global selector when roles have mixed assignments
  • Click a model in the global selector — verify all per-role dropdowns update, toast appears, and per-task section auto-expands
  • Customize individual role models — verify global selector becomes unhighlighted
  • Switch API providers — verify tier defaults apply correctly to each role
  • Run agent sessions and verify the correct model is used per task type

🤖 Generated with Claude Code

CaitlynByrne and others added 2 commits February 8, 2026 20:03
@CaitlynByrne @claude
feat: add per-task-type model assignment with three-tier system
55e556f 
Allows assigning different AI models to different task types (initializer,
coding, testing, spec creation, expansion, assistant, log review). Each
provider defines three tiers (high/mid/low) with smart defaults -- e.g.,
Opus for initialization, Sonnet for coding, Haiku for testing. Users can
override per-role in Settings, and the assistant chat panel gets an
on-the-fly tier switcher.

- registry.py: ROLE_TIER_MAP, provider tiers, get_model_for_role()
- parallel_orchestrator.py: per-role model resolution in spawn methods
- Chat sessions: use get_model_for_role() per session type
- Settings API: read/write per-role overrides, expose tiers
- UI: collapsible "Model per Task" section, assistant tier toggle

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@CaitlynByrne @claude
ui: global model selector sets all role assignments with toast feedback
304fec5 
Clicking a model in the global Model selector now sets all 7 per-role
model assignments to that model, shows an inline toast confirming the
change, and auto-expands the per-task section for further customization.
The global selector only highlights a model when all roles match it.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@CaitlynByrne