Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/sounio-lang/sounio/model-dispatchnpx skills add Sounio-lang/sounio --skill model-dispatchgit clone --depth 1 https://github.com/Sounio-lang/sounioWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sounio-lang/sounio/model-dispatch)<a href="https://agentmods.dev/skills/sounio-lang/sounio/model-dispatch"><img src="https://agentmods.dev/badge/skills/sounio-lang/sounio/model-dispatch.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00047 | $0.00872 |
| Opus 5 | $0.00023 | $0.00436 |
| Sonnet 5 | $0.00009 | $0.00174 |
| Haiku 4.5 | $0.00005 | $0.00087 |
Grade A, and why
model-dispatch scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Sounio Model Dispatch Guide
Workflow
- Identify the task type from the Decision Flowchart or Quick Reference table below.
- Assign the task to the indicated model.
- If the result is uncertain or the first attempt fails, escalate via the Fallback Chain.
- For multi-file compiler changes that touch formal semantics, enable Multi-Model Review.
Decision Flowchart
- Touches 3+ files across compiler stages? -> Opus 4.6
- Requires formal reasoning / soundness? -> Opus 4.6
- Long autonomous run or prose writing? -> GPT-5.3-Codex
- Whole-codebase sweep or impact analysis? -> Gemini 3.1 Pro
- System-level debugging or cross-check? -> GLM-5
- Render example completion (self-contained, no imports)? -> Sonnet 4.6
- Touches self-hosted frontend (check/resolve) + correctness proof? -> Opus 4.6
- PGO pass (IR module, new opcodes, self-test)? -> Opus 4.6
- Website/docs/skills update? -> Sonnet 4.6
- Native backend single-file pass (peephole/frame/encode)? -> Sonnet 4.6
- Multi-file native pipeline change (codegen+lower_ir+encode)? -> Opus 4.6
- Everything else -> Sonnet 4.6
Quick Reference
| Question | Model |
|---|---|
| Change the type checker | Opus 4.6 |
| Write a test | Sonnet 4.6 |
| Write a paper section | GPT-5.3-Codex |
| What will break if I change this? | Gemini 3.1 Pro |
| Why is the bootstrap failing? | GLM-5 |
| Refactor one stdlib file | Sonnet 4.6 |
| Add keyword to lexer+parser+checker+HLIR | Opus 4.6 |
| Second opinion on checker logic | GLM-5 |
| Name a new language feature | GPT-5.3-Codex |
| Find every use of X in 30 files | Gemini 3.1 Pro |
| Complete a render example (heatmap, DAG, 3D) | Sonnet 4.6 |
| Add rasterize_* primitive to stdlib/render/ | Sonnet 4.6 |
| Fix self-hosted --check failure | Opus 4.6 |
| Debug --native-compile bootstrap crash | Opus 4.6 + GLM-5 cross-check |
| Add a new PGO pass (new ir/*.sio module) | Opus 4.6 |
| Run sprint gate and triage failures | Sonnet 4.6 |
| Generate website SVG previews | Sonnet 4.6 |
| Add a new peephole pattern to peephole.sio | Sonnet 4.6 |
| Wire a new native pass in compile_ir_function | Opus 4.6 |
| Debug native binary crash / ELF reloc error | Opus 4.6 + GLM-5 cross-check |
| Add emit_* helper to encode.sio | Sonnet 4.6 |
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 70 lines · 47 tokens per session scan A 57c47acdeaa0
model-dispatch is a skill published in the GitHub repository Sounio-lang/sounio (6 stars, last pushed today), licensed Apache-2.0. It adds 47 tokens to every session and 872 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
vera-language
Write programs in the Vera programming language. Use when asked to write, edit, debug, or review Vera code (.vera files). Vera is a statically typed, purely functional language with algebraic effects, mandatory contracts, and typed slot references (@T.n) instead of variable names.
vera-language
Write programs in the Vera programming language. Use when asked to write, edit, debug, or review Vera code (.vera files). Vera is a statically typed, purely functional language with algebraic effects, mandatory contracts, and typed slot references (@T.n) instead of variable names.
edict-compiler-dev
How to develop the Edict compiler — architecture, coding patterns, structured error conventions, and how to add new features end-to-end. Use this skill when contributing to the Edict TypeScript codebase, adding new AST node types, implementing new builtins, extending the compiler pipeline, fixing bugs in the…
edict-program-writer
How to write correct Edict JSON AST programs. Covers the AST format, node types, ID conventions, common patterns, and how to interpret and fix structured errors. Use this skill when writing Edict programs, creating example programs, producing JSON ASTs for the Edict compiler, demonstrating language features, or…
edict-testing
How to write and run tests for the Edict compiler. Covers vitest conventions, test directory structure, patterns for testing each pipeline stage, structured error assertions, and coverage expectations. Use this skill when writing tests, fixing failing tests, adding coverage for new features, or understanding the test…
skill-creator
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, update or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.