model-routing SKILL.md skillA

model-routing from moonray/repoprompt-workflows is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

mdr add moonray/repoprompt-workflows/model-routing@git:20261011.a5e82c5

Pins git:20261011.a5e82c5 in mdr.lock. Pin the fingerprint instead.

Audit grade

A, 17 of 17 checks passed

Kind

SKILL.md skill

Repository

moonray/repoprompt-workflows

Path

.agents/skills/model-routing/SKILL.md

Versions recorded

2

Latest version

git:20261011.a5e82c5, committed Oct 11, 2026

First version recorded

Oct 10, 2026

Size

5,112 bytes

Repository stars

11

License

MIT

Use when choosing an oracle model or effort level for spec, plan, or review consultations (ask_oracle, context_builder, RPCE models.* settings), or when re-evaluating the Recommended RepoPrompt Architecture after a new model release, roster change, or pricing change. Maintains the canonical role→model routing table and its refresh discipline.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Model Routing

## Recommended RepoPrompt Architecture

Canonical as of 2026-10-10 (user-ratified; last re-evaluated 2026-10-10). Roles are the stable interface; models are data that re-evaluation replaces.

| Role | Consultations | Model | Effort | Why |
|---|---|---|---|---|
| Context & Discovery | repo analysis, spec/plan generation, context-heavy discovery and drafting | GPT-6.1 Sol | high | Cheapest deep cache reads (~$0.10/M cached input). RPCE planning re-reads large context prefixes, so cached-input price dominates run cost. |
| Review & Refine | cross-checking specs/plans, readiness and clarity verdicts, finding tie-breaks | Claude Opus 5.5 | medium | Communicates clearly and reliably catches logical gaps in long documents. Different model family than the drafter, so blind spots don't compound. |
| Escalation | one stuck sub-task after repeated failed attempts (fix loop, unmet complex constraint) | GPT-6 Astra, or Claude Fable 5.1 | low–medium, or high | Strongest raw problem-solving, spent only on the stuck piece. |

Invariants:

- **Cache-first economics.** Cached-input price, not nominal input price, is the primary cost criterion for the Context & Discovery role.
…

Read the whole file at its exact version.

How to install

mdr add moonray/repoprompt-workflows/model-routing@sha256:ffecac6c1b29a30f

[![mdr](https://markdownregistry.com/badge/art_fq2xt7647yvsx2hu.svg)](https://markdownregistry.com/a/art_fq2xt7647yvsx2hu)

Versions

The registry has recorded 2 versions of this file, between Oct 10, 2026 and Oct 11, 2026, more than one a day.

Audit of the latest version

Source

API

GET https://markdownregistry.com/api/v1/artifacts/art_fq2xt7647yvsx2hu
GET https://markdownregistry.com/api/v1/resolve?ref=moonray/repoprompt-workflows/model-routing
GET https://markdownregistry.com/api/v1/blob/ffecac6c1b29a30f339e9579dc625d44d8ff8bc84bbf2c729d18c5680c7660af

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from moonray/repoprompt-workflows

Every file in moonray/repoprompt-workflows

Other files named model-routing

Browse by kind, by grade A, or by owner.