chore(claude): tier model pins across commands, agents, and templates - #31
Conversation
Every slash command was pinned to claude-opus-4-6 (or had no model pin at all), so every turn ran at the highest price regardless of task complexity. Pin by tier instead: sonnet for component specialists and test runners, opus for orchestration/PR review/CI debugging. Aliases instead of full model IDs so pins track the latest model per tier. CLAUDE.md now documents the command table and tier convention so new commands start on the right tier.
Read-only planning command pinned to fable: investigates via cheap subagent fan-out, designs on the top model, and emits a self-contained artifact — a GitHub issue (feeds /lfg <number>) or a docs/plans/ markdown. Plan contract: context block with file paths, verification gates with expected outputs, explicit out-of-scope boundaries, so an executor needs none of the planning session's context. Carves the one fable-pin exception into the tier convention docs.
|
Warning Review limit reachedYou’ve reached a temporary PR review limit under our Fair Usage Limits Policy. Next review available in: 47 minutes Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available. How can I continue?After more reviews become available, a review can be triggered using the To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews. How do review limits work?CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability. For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window. Please refer docs for additional details. Review details⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (12)
Comment |
1 participant
Summary
Replaces the blanket
model: claude-opus-4-6pin on every Claude Code command with a tier-based convention, adds frontmatter to commands that were missing it, and introduces a new/plancommand that puts the expensive model where it pays off.fable/plan(new)docs/plans/markdown for cheaper models to executeopus/lfg,/review-pr,/github-review-comments,/github-ci-failuressonnet/add-component,/audit-components,/check-component,/fix-docs-tests,/tdd,/test-all/plan (second commit)
Read-only planning command pinned to Fable: the capable model audits and writes the spec; cheaper models execute it. Each plan is self-contained — context block with file paths, ordered steps mapped to layer specialists, verification gates with exact commands and expected outputs, explicit out-of-scope boundaries — so
/lfg <issue-number>can execute without any of the planning session's context.Workflow:
/plan add bulk CSV re-import(Fable) →/lfg 1234(Opus) → specialists (Sonnet).Test plan
grep -rn "claude-opus-4" .claude/ CLAUDE.md— no full model IDs remainmodel:tier alias in frontmatter/planregisters as an available skill/tdd) and confirm the turn runs on Sonnet/planon a small feature and execute the resulting issue with/lfgDocs/config only — no application code, no specs affected.