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/kyh/vibedgames/agents-mdgit clone --depth 1 https://github.com/kyh/vibedgamesWrote 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/kyh/vibedgames/agents-md)<a href="https://agentmods.dev/instructions/kyh/vibedgames/agents-md"><img src="https://agentmods.dev/badge/instructions/kyh/vibedgames/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 | $0.02993 | $0.02993 |
| Opus 5 | $0.01496 | $0.01496 |
| Sonnet 5 | $0.00599 | $0.00599 |
| Haiku 4.5 | $0.00299 | $0.00299 |
Grade A, and why
vibedgames AGENTS.md scanned grade A with 1 finding 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 4d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Liveness: `curl -s -o /dev/null -w '%{http_code}' localhost:5173/auth/login` → `200`. (`/` answers `307` — it redirects to a featured game — so don't use it as a health check.) How it starts
The opening of the file, as written. The whole thing — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
vibedgames is an infrastructure platform for deploying, hosting and adding multiplayer to browser games: build a game locally, vg deploy it, and it is served at {slug}.vibedgames.com. The stack is a pnpm/Turborepo monorepo — TanStack Start (React 19 + Vite SSR) on Cloudflare Workers, oRPC + Drizzle on D1, better-auth, PartyServer on Durable Objects, R2 for game bundles.
This is the tool-agnostic guide for coding agents, and it is meant to be run, not just read. Claude also reads CLAUDE.md (product context, architectural decisions, conventions); both point back here.
Quickstart (headless)
pnpm install
cp .env.example .env # every credential, CLI and Worker alike (see Environment)
pnpm dev:web # ← run once, then stop it
pnpm db:local # db:push + seed dev identities
pnpm dev:web # http://localhost:5173
The double dev:web is not a typo. packages/db/drizzle.config.local.ts resolves the Miniflare SQLite file out of apps/web/.wrangler/state/v3/d1/miniflare-D1DatabaseObject/, and that directory only exists after the dev Worker has booted once. Skip the first run and pnpm db:local dies with Local D1 not found. Run 'pnpm dev:web' once to initialize it.
Liveness: curl -s -o /dev/null -w '%{http_code}' localhost:5173/auth/login → 200. (/ answers 307 — it redirects to a featured game — so don't use it as a health check.)
There is no bootstrap script; the five commands above are the provisioning steps. pnpm db:local is idempotent — re-run it any time, and after every schema change.
Environment
One file, two consumers. Everything lives in the repo-root .env; template is .env.example.
| Consumer | Reaches it via | Holds |
|---|---|---|
| drizzle-kit, the wrangler CLI | process.env, loaded by each package's with-env (dotenv-cli) |
CLOUDFLARE_ACCOUNT_ID / _DATABASE_ID / _D1_TOKEN / _API_TOKEN |
the dev Worker, via the Cloudflare env binding |
secrets.required in apps/web/wrangler.jsonc |
BETTER_AUTH_SECRET, R2_*, FAL_API_KEY |
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.
- 4d ago First seen · 155 lines · 2,993 tokens per session scan A fa3775614ccc
vibedgames AGENTS.md is an instructions file published in the GitHub repository kyh/vibedgames (55 stars, last pushed yesterday), licensed MIT. It adds 2,993 tokens to every session, about $0.0150 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
supabase copilot-instructions.md
Copilot instructions for supabase/supabase, covering copilot code review instructions, review policy — read this first, confidence threshold, what not to comment on and what to comment on (priority order).
core CLAUDE.md
Claude Code instructions for module-federation/core, covering claude task parallelization & development guidelines, core principles, maximum parallelization strategy, universal parallel task patterns and for any codebase task.
coss AGENTS.md
AGENTS.md instructions for cosscom/coss, covering agents.md and main documentation.
storefront AGENTS.md
AGENTS.md instructions for saleor/storefront, covering saleor paper — agent guide (always-on), how to get context (read in this order; stop when answered), precedence, critical commands and non-negotiable rules.
pip-web CLAUDE.md
Claude Code instructions for playpip/pip-web, covering claude.md — agent guide for pip, read first (by task), the three layers (respect the boundaries), non-negotiables and commands / definition of done.
whatsapp-automation CLAUDE.md
Instructions for sumioshi/whatsapp-automation, covering vocabulário de acompanhamento (desambiguação), recipe do monitor de mensagem (sem zumbi), modo de envio por chat (confirmar vs autônomo), mídia na nuvem (midiapendente) — não é erro and grupo renomeado ≠ chat morto (slug é estável).