LLM Council on Fireworks AI
dair-ai/dair-academy-plugins
Has several open-weight models answer a question, rank each other's anonymized answers, then lets a chairman model write the final response through Fireworks AI.
9-tier model routing system with cascading classifier fallback and result auto-evaluation
Install Model Routing
$ npx skills add alinaqi/maggy --skill model-routing -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install alinaqi/maggy model-routing --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/model-routing .claude/skills/model-routing && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .claude/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/alinaqi/maggy/tree/main/skills/model-routingType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add alinaqi/maggy --skill model-routing -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install alinaqi/maggy model-routing --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/model-routing .agents/skills/model-routing && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .agents/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add alinaqi/maggy --skill model-routing -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install alinaqi/maggy model-routing --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/model-routing .cursor/skills/model-routing && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .cursor/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/alinaqi/maggy.git --path skills/model-routing--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add alinaqi/maggy --skill model-routing -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install alinaqi/maggy model-routing --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/model-routing .gemini/skills/model-routing && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .gemini/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install alinaqi/maggy model-routingInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add alinaqi/maggy --skill model-routing -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/model-routing .github/skills/model-routing && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .github/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add alinaqi/maggy --skill model-routing -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install alinaqi/maggy model-routing --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/alinaqi/maggy.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/model-routing .opencode/skills/model-routing && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "model-routing" agent skill from https://github.com/alinaqi/maggy/tree/main/skills/model-routing into .opencode/skills/model-routing/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "model-routing", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
model-routing- GitHub stars
- 707
- Token cost
- ~1.5k tokens
- SKILL.md length
- 472 words
- Files
- 1
- Skills in repo
- 71
- Repo updated
- First seen
- Licence
- MIT
At a glance
9-tier model routing system with cascading classifier fallback and result auto-evaluation
- Works in 3 steps: Which model should handle this? — 9-tier… → Is the classifier itself working? —… → Can we verify the result? — Tool-level…
- Tasks that involve Model routing and gateways
- SKILL.md covers How Routing Decisions Are Made, 9-Tier Routing Table, Delegation Commands and Delegation Script Contract, plus 5 more sections
- Calls gemini, claude and codex; needs DEEPSEEK_API_KEY and GEMINI_API_KEY
What it does
Model Routing is an agent skill from alinaqi/maggy. 9-tier model routing system with cascading classifier fallback and result auto-evaluation
Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in AI & LLM Engineering, covering Model routing and gateways. It works with Kimi, DeepSeek and Qwen. The repository describes itself as: What started as an opinionated Claude Code setup kit is now an autonomous AI engineering command center. The licence is MIT.
When your agent uses it
- Tasks that involve Model routing and gateways
Example prompts
- “/model-routing”
Requirements
- A credential in DEEPSEEK_API_KEY
- A credential in GEMINI_API_KEY
Workflow steps
3 steps, taken from the first numbered list in SKILL.md.
- Which model should handle this? — 9-tier cost/complexity classification
- Is the classifier itself working? — Cascading fallback (qwen3 → kimi → deepseek → cache)
- Can we verify the result? — Tool-level fallback + auto-evaluation
What it can do on your machine
Read from SKILL.md and the folder at commit 72a456e. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
geminiclaudecodexollama
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names these keys or tokens, usually read from environment variables:
DEEPSEEK_API_KEYGEMINI_API_KEYOPENAI_API_KEY
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Model Routing loads about 1.5k tokens when it runs. Until then it costs about 26 tokens; SKILL.md has 472 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from alinaqi/maggy at commit 72a456e, republished under its MIT licence (© alinaqi). 472 words, ~1,508 tokens.
Download SKILL.mdSave it as
.claude/skills/model-routing/SKILL.md (or your agent's skills folder).- name
- model-routing
- description
- 9-tier model routing system with cascading classifier fallback and result auto-evaluation
- when-to-use
- When configuring or debugging how prompts are classified and routed to the cheapest capable model
- user-invocable
- false
- effort
- medium
Model Routing System
How Routing Decisions Are Made
Every user prompt goes through a 9-tier classification pipeline before any AI model processes it. The system answers three questions:
- Which model should handle this? — 9-tier cost/complexity classification
- Is the classifier itself working? — Cascading fallback (qwen3 → kimi → deepseek → cache)
- Can we verify the result? — Tool-level fallback + auto-evaluation
The Pipeline
User types prompt
↓
UserPromptSubmit hook fires (~/.claude/hooks/route-task-hook)
↓
Classifier: qwen3 (local, free) classifies into tier
↓ (fails?)
Classifier: kimi (local, free) retries
↓ (fails?)
Classifier: deepseek-flash (~$0.0001) retries
↓ (fails?)
Classifier: cached tier from last success
↓
Hook injects routing decision into Claude's context
↓
Claude delegates to the right model or handles directly9-Tier Routing Table
| Tier | Model | Input (per M) | Output (per M) | Handles |
|---|---|---|---|---|
| 0 | Qwen3 (local) | $0 | $0 | grep, find, shell, syntax, log reading |
| 1 | Gemini 2.5 Flash-Lite | $0.10 | $0.40 | Bulk extraction, classification, CIG pipelines |
| 2 | DeepSeek V4 Flash | $0.14 | $0.28 | Simple code, CRUD, test writing, small fixes |
| 3 | DeepSeek V4 Pro | $0.44 | $0.87 | Multi-file features, refactors, debugging (~80% of work) |
| 4 | Gemini 2.5 Flash | $0.15 | $0.60 | Multimodal (images, video, audio), brand analysis |
| 5 | Kimi K2.6 | $0.60 | $2.50 | Code review, commit messages, diff summaries |
| 6 | Gemini 3.1 Pro + Search | $1.25 | $10.00 | Deep research, Google grounding, 2M context |
| 7 | Codex | varies | varies | Bulk generation, code review |
| 8 | Claude Sonnet/Opus | $3-5 | $15-25 | Architecture, security, quality-critical |
Delegation Commands
When the hook says "delegate to X", run the matching command and return its output:
# Tier 0 — Qwen3
~/bin/qwen3 "prompt"
# Tier 1 — Gemini Flash-Lite
~/bin/gemini --flash-lite "prompt"
# Tier 2 — DeepSeek Flash
~/bin/deepseek --flash "prompt"
# Tier 3 — DeepSeek Pro
~/bin/deepseek --pro "prompt"
# Tier 4 — Gemini Flash
~/bin/gemini --flash "prompt"
# Tier 5 — Kimi
~/bin/kimi --quiet -p "prompt"
# Tier 6 — Gemini Pro Search
~/bin/gemini --pro-search "prompt"
# Tier 7 — Codex
codex exec "prompt"
# Tier 8 — Claude
# Handle directly (no delegation)Delegation Script Contract
Every ~/bin/ script follows the same pattern:
- Accepts prompt as argument:
script "what is 2+2" - Model flags:
--flash,--pro,--flash-lite,--pro-search - Quiet mode:
--quiet(where applicable) - Output: writes response to stdout, errors to stderr
- Exit codes: 0 on success, non-zero on failure
Available Scripts
~/bin/
├── qwen3 # Shell: curl to local Ollama API
├── kimi # Shell: execs Kimi CLI binary
├── deepseek # Python: httpx to DeepSeek Anthropic-compat API
├── gemini # Python: httpx to Gemini OpenAI-compat API
├── research # Python: multi-backend research with auto-evaluation
└── route-task # Shell: qwen3-powered task classificationClassifier Fallback Chain
The classifier itself can fail. When it does, cascading fallback kicks in:
| Level | Classifier | Cost | Threshold |
|---|---|---|---|
| 1 | qwen3 (Ollama) | $0 | 2s connect, 8s classify |
| 2 | kimi CLI | $0 | Local process |
| 3 | deepseek-flash | ~$0.0001 | API call |
| 4 | Cached tier | $0 | From ~/.claude/routing-cache.json |
The cache (~/.claude/routing-cache.json) saves the last successful tier and timestamp. After compaction, when Ollama may be briefly unreachable, the cache ensures routing continues without dropping to CLAUDE by default.
Show full SKILL.md (158 more words)Show less
Tool Fallback Protocol
When Claude's built-in tools fail, external backends take over:
| Failed Tool | Fallback 1 | Fallback 2 |
|---|---|---|
| WebSearch / WebFetch | ~/bin/research "query" | ~/bin/deepseek --pro "query" |
| Read / file access | cat via Bash | — |
| Grep | grep -r via Bash | — |
Research Tool (~/bin/research)
Multi-backend research with auto-evaluation:
- Tries deepseek-flash → deepseek-pro in sequence
- Scores results 0-10 on content quality, structure, length
- Auto-adjusts preferred backend based on evaluation scores
- View stats:
~/bin/research --eval - Score log:
~/.claude/research-eval.jsonl
Maggy Integration
Maggy's model_router.py mirrors the same 9-tier structure in DEFAULT_TIERS. The PiAdapter uses the same delegation scripts for execution. Task type overrides in routing_rules_defaults.py ensure:
research,competitor→ Gemini Pro Search (Google grounding)bulk→ Gemini Flash-Lite (cheapest)security,architecture,planning→ Claude (quality-critical)docs,tests→ DeepSeek Pro (cost-efficient)review→ Claude (security + architecture depth)
Environment
# Required for delegation scripts (in ~/.zshrc)
export DEEPSEEK_API_KEY="sk-..."
export GEMINI_API_KEY="..." # For gemini delegator
export OPENAI_API_KEY="sk-..." # For codex CLI
# Ollama must be running locally for qwen3
ollama serve # or launch at startupObservability
- Routing log:
~/.claude/routing-log.jsonl— every classification with tier, classifier used, tokens saved - Routing cache:
~/.claude/routing-cache.json— last tier for post-compact recovery - Research eval:
~/.claude/research-eval.jsonl— per-query backend scoring - Maggy routing heatmap: Dashboard → Models tab → per-model reward scores
© alinaqi, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in skills/model-routing of alinaqi/maggy.
Open the folder on GitHubat commit 72a456e
Compare with similar skills
Model Routing next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Model Routing this skillalinaqi/maggy | 707 | — | ~1.5k | Automated safety check: Pass | MIT | |
| LLM Council on Fireworks AIdair-ai/dair-academy-plugins | 614 | — | ~5k | Automated safety check: Notes | MIT | |
| Claude Maintain ModelsKiln-AI/Kiln | 5.2k | — | ~15k | Automated safety check: Notes | Custom licence | |
| LLM Pipeline Profiler AnalysisBBuf/AI-Infra-Auto-Driven-SKILLS | 938 | — | ~3.9k | Automated safety check: Pass | None | |
| Update Ollama Cloud Modelsheypinchy/pinchy | 182 | — | ~3.9k | Automated safety check: Notes | AGPL-3.0 | |
| OmniRoute Chat CLIdiegosouzapw/OmniRoute | 75k | — | ~345 | Automated safety check: Pass | MIT |
Similar skills
Claude Maintain Models
Kiln-AI/Kiln
Add new AI models to Kiln's mlmodellist.py and produce a Discord announcement.
LLM Pipeline Profiler Analysis
BBuf/AI-Infra-Auto-Driven-SKILLS
Breaks LLM torch profiler traces down by forward pass, layer and kernel, with timing tables and Perfetto time ranges for the layers you want to inspect.
Update Ollama Cloud Models
heypinchy/pinchy
A skill your agent uses when a new Ollama Cloud model is announced or available (e.g.
OmniRoute Chat CLI
diegosouzapw/OmniRoute
Sends chat completions, streams responses, and opens an interactive REPL against any OmniRoute-routed model provider.
Nowait Reasoning Optimizer
davila7/claude-code-templates
Implements the NOWAIT technique for efficient reasoning in R1-style LLMs.
More from alinaqi/maggy
All 71 skills in this repoAeo Optimization
alinaqi/maggy
AI Engine Optimization - semantic triples, page templates, content clusters for AI citations
Agent Teams
alinaqi/maggy
Claude Code Agent Teams - default team-based development with strict TDD pipeline enforcement
AI Models
alinaqi/maggy
Latest AI models reference - Claude, OpenAI, Gemini, Eleven Labs, Replicate
Android Java
alinaqi/maggy
Android Java development with MVVM, ViewBinding, and Espresso testing
Android Kotlin
alinaqi/maggy
Android Kotlin development with Coroutines, Jetpack Compose, Hilt, and MockK testing
Autonomous Testing
alinaqi/maggy
AI-driven testing agent that auto-discovers, generates, executes, evaluates, and fixes tests for any project type
Related
Categories
Questions about Model Routing
What does Model Routing do?
9-tier model routing system with cascading classifier fallback and result auto-evaluation. Model Routing is an agent skill from alinaqi/maggy.
When should I use Model Routing?
Model Routing fits situations like: tasks that involve Model routing and gateways.
How do I install Model Routing in Claude Code?
Run `npx skills add alinaqi/maggy --skill model-routing -a claude-code`. Or copy the skill folder (skills/model-routing in alinaqi/maggy) into .claude/skills/model-routing in your project. Claude Code loads it when a task matches its description.
How do I install Model Routing in Codex?
Run `npx skills add alinaqi/maggy --skill model-routing -a codex`. Or copy the skill folder (skills/model-routing in alinaqi/maggy) into .agents/skills/model-routing in your project. Codex loads it when a task matches its description.
Can I use Model Routing in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add alinaqi/maggy --skill model-routing -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/model-routing, .gemini/skills/model-routing, .github/skills/model-routing and .opencode/skills/model-routing in your project.
What does Model Routing need to run?
Going by SKILL.md and its folder, Model Routing needs the command-line tools its instructions call (gemini, claude, codex and ollama) and credentials named DEEPSEEK_API_KEY, GEMINI_API_KEY and OPENAI_API_KEY. Our summary lists: A credential in DEEPSEEK_API_KEY; A credential in GEMINI_API_KEY.
Does Model Routing access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Model Routing safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Model Routing use?
Model Routing is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Model Routing use?
About 1.5k tokens (SKILL.md is roughly 6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Model Routing?
Skills that share tags, products or a category with Model Routing: LLM Council on Fireworks AI (dair-ai/dair-academy-plugins, 614 stars), Claude Maintain Models (Kiln-AI/Kiln, 5.2k stars), LLM Pipeline Profiler Analysis (BBuf/AI-Infra-Auto-Driven-SKILLS, 938 stars) and Update Ollama Cloud Models (heypinchy/pinchy, 182 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Model Routing?
alinaqi (a GitHub user) maintains it in alinaqi/maggy, which has 707 GitHub stars. The repository holds 71 skills in this directory. The repository was last updated on September 24, 2026.
Source: alinaqi/maggy on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.