Skip to content

feat: tier Claude models by task in the upgrade pipeline - #26

Merged
george-elphick-talieisin merged 1 commit into
mainfrom
feat/tiered-models
Sep 21, 2026
Merged

george-elphick-talieisin merged 1 commit into
mainfrom
feat/tiered-models

Conversation

@george-elphick-talieisin

@george-elphick-talieisin george-elphick-talieisin commented Sep 21, 2026 •

Copy link
Copy Markdown
Contributor

Summary

  • Every Claude job defaulted to fable. The model now follows task difficulty: round 1 keeps fable in upgrade mode and uses opus for rescue and suggest; round 2 (verifying Codex findings) defaults to opus; Copilot responses default to sonnet.
  • model stays an override for both rounds (its default is now empty). A new review-model input overrides round 2 alone (review-model || model || 'opus'). No template caller sets either, so the new defaults apply org-wide on merge, and any hand-set override keeps working.
  • Codex reviewed the plan and raised one finding (round-2 override precedence), which is fixed. The companion docs and security-fix change are in a Talieisin/automation PR; merge this one first.

Test plan

  • YAML parses; actionlint shows no new findings compared with main
  • Next upgrade-mode run: claude-code-action step shows --model fable
  • Next rescue/suggest run and any round 2: --model opus
  • Next Copilot-response run: --model sonnet

View with [code]smith Autofix with [code]smith
Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.

Round 1 keeps fable for upgrade mode and uses opus for rescue and
suggest; round 2 defaults to opus via a new review-model input (model
still overrides both rounds). Copilot responses default to sonnet.
Copilot AI lite review requested due to automatic review settings September 21, 2026 09:38

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot review overview

🟢 Approval recommended

No unresolved blocking issues were identified.

Review effort: Lite
Findings: None

What changed in this PR

Updates Claude model selection in upgrade and Copilot workflows with task-appropriate defaults while preserving overrides.

Changes:

  • Adds per-mode and per-round model defaults.
  • Adds a round-two review-model override.
  • Changes Copilot responses to default to sonnet.
File Description
.github/​workflows/​dependabot-upgrade.yml Implements per-round and per-mode model selection.
.github/​workflows/​copilot-response.yml Sets sonnet as the default model.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

@george-elphick-talieisin
george-elphick-talieisin merged commit 99e4a38 into main Sep 21, 2026
2 checks passed
@george-elphick-talieisin
george-elphick-talieisin deleted the feat/tiered-models branch September 21, 2026 09:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants