# osolmaz/switch

> A TypeScript framework for building harnesses around decision models: your code turns task state into typed questions, and the model picks answers step by step until it can respond. Aimed at developers building eval judges, game agents and other finite-choice loops.

- Magnitude: 6.1 out of 10 — Early signal
- Stars: 183 total · +27 stars measured 2026-10-30, 05:51–07:57 UTC, ≈ 121 by evening
- Star trust: star growth looks organic
- Category: AI agents · Language: TypeScript · License: MIT · Created: 2026-10-25 · Last push: 2026-10-30
- GitHub: https://github.com/osolmaz/switch · Page: https://gitnova.dev/en/r/osolmaz/switch

## Useful for

- Build a rubric judge that reads only the needed pages and images of a document
- Run a chess domain against Stockfish or random moves via the CLI
- Replay a saved trace to inspect states and answers without calling a model

## Why it’s here

- Star-counter measurements on 2026-10-30 (UTC), 05:51–07:57: 156 → 183 stars (+27). This is the change over that interval.
- Estimated end-of-day forecast: about +121 stars, using observed gains and the previous day.
- The repository is 5 days old and already has 183 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Top new repositories this week: #178.
- About 37 forks a day — people are taking the code.
- Recent forks include notable developers: @lgs (296 followers).

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 18
- Issues and pull requests: 42
- Watchers: 0
- Average over the last week: 49 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 10
- Latest release: v0.1.1 (2026-10-29)

## Stars per day, last 6 days (oldest → newest, today is partial)

2026-10-25 … 2026-10-30: 0, 0, 0, 0, 174, 27

## Spotted in now

- Top new repositories this week: #178

## More in this category

1. **HKUDS/DeepTutor** — 8.9 · Breakout · AI agents · Python · +917 stars measured 2026-10-30, 00:15–07:57 UTC, ≈ 2,492 by evening
   A lifelong personalized AI tutor with long-term memory: builds learning paths, reads documents and books, and runs practice and assessments via a multi-agent system and RAG.
   Full card: https://gitnova.dev/en/r/HKUDS/DeepTutor.md
2. **ccch1mneyyy/dsh-go** — 8.4 · Early signal · AI agents · Dart · +888 stars measured 2026-10-30, 00:18–07:57 UTC, ≈ 1,811 by evening
   Unofficial Android client for DeepSeek Harness: runs DSH on the phone in a built-in Linux environment without root or Termux, and can also remotely control DSH on your computer.
   Full card: https://gitnova.dev/en/r/ccch1mneyyy/dsh-go.md
3. **Donchitos/Claude-Code-Game-Studios** — 8.1 · Breakout · AI agents · Shell · +516 stars measured 2026-10-30, 00:16–07:57 UTC, ≈ 1,325 by evening
   A Claude Code configuration that turns a single session into a structured game studio: 49 specialized subagents, 74 slash-command skills, hooks, rules and templates mirroring a real studio hierarchy. It adds design reviews, QA passes and…
   Full card: https://gitnova.dev/en/r/Donchitos/Claude-Code-Game-Studios.md
4. **milind-soni/OpenMausBot** — 7.9 · Breakout · AI agents · TypeScript · +293 stars measured 2026-10-30, 00:15–07:57 UTC, ≈ 849 by evening
   ,
   Full card: https://gitnova.dev/en/r/milind-soni/OpenMausBot.md
5. **pbakaus/impeccable** — 7.8 · Breakout · AI agents · Rust · +729 stars measured 2026-10-30, 00:15–07:57 UTC, ≈ 2,284 by evening
   A set of commands and detector rules for AI coding agents that helps them produce better frontend design instead of generic templates. Installs as a skill into Claude Code, Cursor, Codex and other harnesses.
   Full card: https://gitnova.dev/en/r/pbakaus/impeccable.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-10-30 08:09 UTC, updated every 30 minutes.