agentleFS
Sign inStart free
gigio1023gigio1023/agent-skills/skills/minimal/model-routing/SKILL.md

model-routing

Skill2 starsChanged 10 days agoMITView on GitHub

Choose each subagent's model and reasoning effort and state the resolved settings before spawning, whenever a Claude Fable lead delegates in a harness with subagents or a GPT-6 lead spawns subagents or a GPT worker is assigned, including inside codex-delegate missions. Also installs or maintains the shipped subagent definitions, GPT worker defaults, and Codex role files, and brings Fable in as a judge under another lead when asked. Use for "which model for this subagent", "set the workers' effort", "subagent model 정해줘", "worker 모델 골라줘". NOT for deciding whether to delegate (orchestrate-subagents) or launching an external mission (codex-delegate).

30 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ecosystem-reference, fractal-delegation and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • naman10parikhSkill

    ecosystem-reference

    Quick-access directory of Claude Code ecosystem repos, skills, plugins, MCPs, and patterns. Use when improving the harness.

    Also in 5 projects that use this

  • naman10parikhSkill

    fractal-delegation

    Also in 5 projects that use this

  • KirchliveSkill

    skill-routing

    Decides which of the installed skills a task needs and loads them before the work starts. This skill should be used at the beginning of any multi-step task — build, fix, refactor, migrate, review, plan, release, research, write, design, video — and again at every phase change, after context compaction, and when work moves into a different part of the codebase. Use it for requests like "let's build X", "fix this bug", "refactor Y", "plan the migration", "review this", and also when the user asks which skill applies, why a skill did not fire, or asks to check what is available.

    Also in 5 projects that use this

  • sublimotionSkill

    agent-runtime

    Launch and manage detached agent runs on EKS via the agent-runner CLI. Use when the user says "launch an agent run", "agent runtime", "agent-runner", "run this spec on the cluster", "check run status", "agent run logs", or "stop a run". Do NOT use for writing the agent-runner code itself (edit the sibling ../agent-runner repo), Terraform of the foundation (use terraform-automation), or GPU serving deployments (use infra-deployer).

    Also in 5 projects that use this

  • GoZippySkill

    repo-scout

    Also in 5 projects that use this

  • BenjaminGSkill

    harness-review

    Turn one PR's review comments into harness fixes — classify each cause by the most solid place it can live (codebase, lint, rules, skill), then draft the fixes in a separate PR.

    Also in 5 projects that use this

  • +6 more from naman10parikh

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 28 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the lead's identity from what the harness states, such as the system prompt or the model setting.

    line 13
  2. Distinctive:

    When maintaining this skill, edit only the family whose facts or policy changed.

    line 38

“Common” is measured: kept by at least 6 in 10 of the 28 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Model Routing

This skill ships two routing policies, one per model family. They share orchestration principles but not model tables, default routes, or effort rules, so load only the policy that applies and keep each family's rules on that family's models.

Select the Policy

Read the lead's identity from what the harness states, such as the system prompt or the model setting. The harness alone does not settle it: Claude Code can reach GPT models through a proxy, and a proxied Fable lead can run inside Codex.

LeadLoad
Claude Fable 5 or 5.1; Mythos 5 and 5.1 count as Fablereferences/fable/policy.md
GPT-6 Astra, GPT-6.1 Sol, or legacy GPT-6 Sol, including a lead inside a codex-delegate missionreferences/gpt/policy.md
Any other model, or a lead that cannot be establishedNo shipped policy for that lead: say so and use the harness defaults, apart from the cross-family cases below. Treat an unestablished lead as non-Fable.

Two cases open a policy other than the lead's. A worker from another family follows its own family's policy, for that worker only. Installing or maintaining shipped files follows the policy of the family whose files change, whoever edits them: a GPT lead updating Fable agent definitions reads the Fable policy and its harness adapters. GPT workers follow references/gpt/policy.md under any lead, including a Fable lead or a lead with no shipped policy. Under a non-Fable lead, the Fable policy applies only on request, when the user names it (formerly fable5-model-routing) or asks to put Fable on the judgment; its "Another model leads" rule in references/fable/policy.md governs that case. Without such a request, Anthropic workers under a non-Fable lead have no shipped policy and use the harness defaults.

Never carry one family's tier table, default routes, effort defaults, or role files to the other family's models. For example, Fable workers inherit the selected session effort even though the GPT roles ship at xhigh, and the GPT division between ordinary Sol research and Astra judgment does not govern Sonnet or Haiku routes.

Boundaries

This skill selects model and effort for a delegation that the user or the lead has already decided on. An explicit user instruction about delegation, model, effort, or budget governs; otherwise orchestrate-subagents and the lead's judgment decide whether to delegate. Install or update subagent definitions, worker defaults, or role files only under the user's configuration grant. Each policy states these boundaries in full for its family.

Family Files

Open a family's policy first; it says when to read the rest.

FamilyPolicyHarness mechanicsMaintenance notesInstallable assets
Fablereferences/fable/policy.mdreferences/fable/harness-adapters.mdreferences/fable/source-notes.mdassets/fable/agents/claude-code/
GPTreferences/gpt/policy.mdreferences/gpt/harness-adapters.mdreferences/gpt/source-notes.mdassets/gpt/codex/

When maintaining this skill, edit only the family whose facts or policy changed. The policies are not mirrored, so a change to one does not imply the same change to the other.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free