claude

model-router

Automatic model routing — thinking stays in the main session; building goes to parallel sonnet builders (a coordinator splits big plans); git, pipelines, cloud and infra commands go to a background haiku operator; diagnosis goes to a sonnet senior-operator.

1 on GitHubby Ramon de la Rosav1.9.0updated 1 month ago
/

Plugin model-router from marketplace model-routing in rdelaros/my_claude_model_routing. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (2)

  • setup

    Configure and check the model-router: the model each tier uses (haiku, sonnet, opus, fable), whether those models answer on this provider, and whether the router is active (doctor). Use after installing, when a builder or operator fails with a model error, or when asked whether the router is working

  • train

    Training mode for the model-router: explain every routing decision, ask the user whether it is right before acting, and record corrections. Use when the user wants to learn how the router works, tune it, or asks to train it. Arguments: quiet, review, off.

Subagents (4)

  • builder

    Implements one approved work package (edits, new files, scoped tests) and reports files, tests and blockers. No design, commits or refactors.

  • coordinator

    Splits an approved multi-file plan into work packages, runs builders in parallel, runs the full test suite, reports once. Not for design or one-builder plans.

  • operator

    Background runner for known commands: git, PRs/MRs, pipelines, az, terraform plan, kubectl, jobs, queries, Jira. Reports in under 100 words; never edits code.

  • senior-operator

    Background diagnosis: why a pipeline, plan, job or deployment failed; multi-step investigations; conflicted rebases. Reports cause and next step; never edits code.

Hooks (11)

  • SessionStart

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.py" rules || python "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.py" rules || true

  • SessionStart

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.py" corrections || python "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.py" corrections || true

  • UserPromptSubmit

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/ensure-rules.py" rules || python "${CLAUDE_PLUGIN_ROOT}/hooks/ensure-rules.py" rules || true

  • UserPromptSubmit

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/ensure-rules.py" corrections || python "${CLAUDE_PLUGIN_ROOT}/hooks/ensure-rules.py" corrections || true

  • UserPromptSubmit

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/context-watch.py" || python "${CLAUDE_PLUGIN_ROOT}/hooks/context-watch.py" || true

  • UserPromptSubmit

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/ops-nudge.py" reset || python "${CLAUDE_PLUGIN_ROOT}/hooks/ops-nudge.py" reset || true

  • PreToolUse

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/agent-model.py" || python "${CLAUDE_PLUGIN_ROOT}/hooks/agent-model.py" || true

  • PreToolUse

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/guard-writes.py" || python "${CLAUDE_PLUGIN_ROOT}/hooks/guard-writes.py" || true

  • PostToolBatch

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/ops-nudge.py" count || python "${CLAUDE_PLUGIN_ROOT}/hooks/ops-nudge.py" count || true

  • SubagentStart

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/agent-context.py" || python "${CLAUDE_PLUGIN_ROOT}/hooks/agent-context.py" || true

  • SubagentStop

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/report-budget.py" || python "${CLAUDE_PLUGIN_ROOT}/hooks/report-budget.py" || true

View source on GitHub

Similar plugins

Token usage tracking, model cost attribution per agent, budget alerts, and optimization recommendations

24 skills · 1 command · 1 agent · hooks

ruvnet/ruflo 81,589

aws-core

claudecodex

Build, deploy, and operate applications on AWS. Skills to author infrastructure-as-code, use core services, and complete common tasks.

27 skills · 1 MCP server

anthropics/claude-plugins-official 47,076

agenthub

claude

Multi-agent collaboration — spawn N parallel subagents that compete on code optimization, content drafts, research approaches, or any task that benefits from diverse solutions. 7 slash commands (/hub-init, /agenthub:spawn, /hub-status, /agenthub:eval, /agenthub:merge, /agenthub:board, /hub-run), agent templates, DAG-based orchestration, LLM judge mode, message board coordination.

8 skills · 1 agent · hooks

alirezarezvani/claude-skills 29,725

Essential development subagents for everyday coding tasks - backend, frontend, fullstack, mobile, and API design

VoltAgent/awesome-claude-code-subagents 26,513

PowerPoint generation and editing plugin with text-only QA (no vision required). Uses subagents for cover, TOC, content, section dividers, summary slides, and template-based PPT editing workflows.

5 skills · 5 agents

MiniMax-AI/skills 13,828

Triages a repository's open GitHub issues and pull requests via the gh CLI: optionally merges ready bot and maintainer-approved PRs and spawns review subagents for unreviewed ones, closes already-resolved issues with referenced explanations, cross-links issues with pending fix PRs, and assigns local-only priority and change-size estimates.

1 skill

trailofbits/skills 8,166

Maintain model-router?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted