Skip to content

feat(model-overlays): add opus-4-8 overlay + resolver mapping#2243

Open
chrisquorum wants to merge 1 commit into
garrytan:mainfrom
chrisquorum:feat/opus-4-8-overlay
Open

feat(model-overlays): add opus-4-8 overlay + resolver mapping#2243
chrisquorum wants to merge 1 commit into
garrytan:mainfrom
chrisquorum:feat/opus-4-8-overlay

Conversation

@chrisquorum

Copy link
Copy Markdown

What

claude-opus-4-8 currently resolves through models.ts to the generic claude
overlay. This adds a dedicated opus-4-8 overlay so 4.8 sessions inherit the
Opus-4.x family nudges — the same treatment opus-4-7 already gets:

  • Effort-match the step — quick on mechanical work, deep reasoning reserved for hard subproblems.
  • Pace questions to the skill — one question per turn when the skill carries STOP. AskUserQuestion; batching becomes the explicit exception.
  • Literal interpretation awareness — deliver the full scope asked, not just the first/obvious item.

Changes

  • model-overlays/opus-4-8.md (new) — {{INHERIT:claude}} + the three nudges.
  • scripts/models.ts — add opus-4-8 to ALL_MODEL_NAMES and a resolveModel
    mapping (/^claude-opus-4-8(-|$)/ → opus-4-8), placed before the generic claude fallback.
  • test/model-overlay-opus-4-8.test.ts (new) — gate-tier test mirroring model-overlay-opus-4-7.test.ts.

Notes

  • Content mirrors opus-4-7's validated family nudges — this is a family-inheritance
    overlay, not empirically re-tuned for 4.8. Retune as 4.8-specific patterns emerge.
  • preflight-agent-sdk.ts's paid-eval reference model intentionally left at opus-4-7.

Validation

  • bun test test/model-overlay-opus-4-8.test.ts → 9 pass, 0 fail.
  • resolveModel('claude-opus-4-8') and 'claude-opus-4-8-20260101'opus-4-8; opus-4-7 and claude-sonnet-* unchanged.
  • bun run gen:skill-docs → no generated-doc churn (only the 3 files above change); the docs-freshness CI gate stays green.
  • bun run typecheck → no new errors in the changed files.

Give claude-opus-4-8 a dedicated overlay instead of the generic claude
fallback: effort-matching, skill-paced questions, complete-scope literal
execution. Mirrors opus-4-7's family nudges. Adds the models.ts resolver
mapping and a gate-tier test mirroring model-overlay-opus-4-7.test.ts.
@trunk-io

trunk-io Bot commented Jul 12, 2026

Copy link
Copy Markdown

Merging to main in this repository is managed by Trunk.

  • To merge this pull request, check the box to the left or comment /trunk merge below.

After your PR is submitted to the merge queue, this comment will be automatically updated with its status. If the PR fails, failure details will also be posted here

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.

1 participant