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 instructions/tofunori/atelier-studio/agents-mdgit clone --depth 1 https://github.com/tofunori/atelier-studioWrote 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/instructions/tofunori/atelier-studio/agents-md)<a href="https://agentmods.dev/instructions/tofunori/atelier-studio/agents-md"><img src="https://agentmods.dev/badge/instructions/tofunori/atelier-studio/agents-md.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 5.1 | $0.03014 | $0.03014 |
| Opus 5 | $0.01507 | $0.01507 |
| Sonnet 5 | $0.00603 | $0.00603 |
| Haiku 4.5 | $0.00301 | $0.00301 |
Grade B, and why
atelier-studio AGENTS.md scanned grade B with 2 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
s'il est nécessaire, se configure séparément dans `~/.codex/config.toml` et Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -s -m 5 -H "x-atelier-token: $TOKEN" "http://127.0.0.1:$PORT/health" # {"ok":true,...} How it starts
The opening of the file, as written. The whole thing — 204 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Protocole de relance d'Atelier Studio (agents : Codex, Claude, forks)
À suivre EXACTEMENT avant de valider toute modification qui touche l'app, le runtime, le build ou la galerie. Aucune improvisation. Une modification limitée aux documents/plans ne nécessite pas de rebuild. Le non-respect crée des zombies (serveurs galerie/sidecar orphelins) qui servent du vieux code et font croire que le fix « ne marche pas ».
Ce qu'il faut savoir (2 min)
- L'app buildée =
src-tauri/target/release/bundle/macos/Atelier.app. Son process s'appelletauri-app(PAS « Atelier ») →pkill -x Atelierne matche jamais. - 3 familles de process : l'app (
tauri-app), le sidecar chat (Rust seul depuis plan 065 phase A :Resources/rust-server/atelier-studio-server; l'ancien repliATELIER_BACKEND=nodeest retiré, voirdocs/soak/033-COMPLETE.md), les serveurs galerie (node …/server/main.mjs, un par projet ouvert — ils SURVIVENT aux relances et l'app les réutilise → zombies = vieux code servi). npm run tauri devne survit PAS lancé par un agent (reaping du harness). Seul Thierry le lance depuis son terminal. Les agents utilisent le BUILD.- Vite dev sert
src/en direct, mais l'app buildée fige tout au build : aucun changement n'est visible sans rebuild. - Le build quotidien génère seulement le bundle macOS
.app. Le DMG est réservé aux releases explicites afin d'éviter les grosses images temporairesrw.*.dmg. - Chaque worktree garde ses propres
target/pour éviter les collisions. Cargo passe parsccachequand il est installé afin d'accélérer une reconstruction après nettoyage ; le wrapper retombe surrustcsinon. - Le protocole se lance depuis le worktree qui contient les changements. Il
détecte sa racine Git et construit/ouvre le
Atelier.appde CE worktree. Une seule instance d'Atelier peut tourner à la fois : la phase d'arrêt reste donc globale, même lorsque le build est local au worktree. - La galerie a 2 copies : source
gallery/(à committer) et bundlesrc-tauri/gallery-dist/(régénéré parscripts/stage-gallery.sh). Modifiergallery/assets/*sans restager = bundle périmé.
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 · 204 lines · 3,014 tokens per session scan B 4e861d1ec2e4
atelier-studio AGENTS.md is an instructions file published in the GitHub repository tofunori/atelier-studio (2 stars, last pushed 2d ago), licensed MIT. It adds 3,014 tokens to every session, about $0.0151 per session on Opus 5. A static security scan graded it B with 2 findings (reads agent configuration directories, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
podium-app CLAUDE.md
Instructions for abrahao-dev/podium-app, covering podium, stack (fixed — do not substitute), pty ipc contract, cross-platform rules and frontend layout.
framework AGENTS.md
AGENTS.md instructions for ai-driven-dev/framework, covering agents.md, behavior, communication, action and memory management.
moonproxy-desktop AGENTS.md
AGENTS.md instructions for MoonProxyHQ/moonproxy-desktop, covering moonproxy, 功能, 技术栈, 目录结构 and 配置与数据存储.
toh-framework CLAUDE.md
Claude Code instructions for wasintoh/toh-framework, covering claude.md — toh framework (repo development guide), what this is, everyday commands, verification protocol and single source, transformed per ide.
hatch3r CLAUDE.md
Claude Code instructions for hatch3r/hatch3r, covering hatch3r — development instructions, architecture, development commands, two-axis pillar framework (2.0.0) and orchestrator self-discipline (bypass protection).
Token-Dashboard CLAUDE.md
Instructions for Arylmera/Token-Dashboard, covering claude.md, project overview, status, architecture and data source.