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/ddyy/minshop/agents-mdgit clone --depth 1 https://github.com/ddyy/minshopWhat 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 | $0.03889 | $0.03889 |
| Opus 5 | $0.01945 | $0.01945 |
| Sonnet 5 | $0.00778 | $0.00778 |
| Haiku 4.5 | $0.00389 | $0.00389 |
Grade A, and why
minshop 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 2d 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.
curl -s localhost:4321/cdn-cgi/explorer/api/d1/database/DB/raw \ How it starts
The opening of the file, as written. The whole thing — 260 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — working on minshop
A guide for contributors and automated tooling making changes to minshop: a small, full-Cloudflare ecommerce store (Astro 7 SSR on Workers + D1 + R2 + Stripe / Lightning). Read this before editing. It's the map, the rules, the recipes, and the traps.
The loop (run this constantly)
nvm use 22 # REQUIRED — the supported toolchain runs on Node 22
npm run verify # complete storefront + D1 + MCP green/red gate
npm run verify is the single signal that a change is sound: it runs unit tests,
full Astro diagnostics, the production build, the clean-room D1 integration, and
the MCP typecheck/deployment dry run. If it's green, the change holds together.
Run it after every meaningful edit, not just at the end.
Other commands: npm run dev (astro dev), npm run preview (wrangler dev =
production mode, for testing middleware/auth), npm run db:migrate (local D1),
npm test.
Inspecting local data: use Local Explorer, not the CLI
Any local dev server (npm run dev, npm run preview) serves Cloudflare's
Local Explorer — a UI at /cdn-cgi/explorer and a self-describing REST API at
/cdn-cgi/explorer/api covering D1, KV, R2, Durable Objects and Workflows.
It reads the same local state the app does, with no setup.
Prefer it over npx wrangler d1 execute --local for reads and fixture setup:
that spawns a process each time (~1.2s) where this is an HTTP call (~0.02s) —
roughly 70x, which is the difference between checking state freely and avoiding
it. GET /cdn-cgi/explorer/api returns the OpenAPI spec, so the endpoints are
discoverable rather than memorized. The D1 database id is the binding name:
curl -s localhost:4321/cdn-cgi/explorer/api/d1/database/DB/raw \
-H 'content-type: application/json' \
-d '{"sql":"SELECT status, COUNT(*) FROM orders GROUP BY status"}'
Local only, and unauthenticated on the dev origin — it is never exposed on a
deployed Worker, but do not bind a dev server to a public interface. Anything
touching --remote still goes through wrangler. The API reports itself as
version 0.0.1, so use it for working loops, not as the basis of a test gate.
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.
- 2d ago First seen · 260 lines · 3,889 tokens per session scan A 91207e421ddb
minshop AGENTS.md is an instructions file published in the GitHub repository ddyy/minshop (120 stars, last pushed 7d ago), licensed MIT. It adds 3,889 tokens to every session, about $0.0194 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
zunfurl AGENTS.md
Instructions for ZUnfurl/zunfurl, covering agents.md, 默认沟通, 项目边界, a/b/c profile and 项目契约.
orbit.sametbasbug.dev AGENTS.md
AGENTS.md instructions for sametbasbug/orbit.sametbasbug.dev, covering agents.md, orbit nedir, beş kural, komutlar and repo haritası.
hands AGENTS.md
AGENTS.md instructions for botiverse/hands, covering hands agent guide, start here, new machine github setup, repository map and workflow rules.
flarestarter AGENTS.md
Instructions for FlareStarter/flarestarter, covering agents.md, what this is, structure, conventions and commands.
microfeed AGENTS.md
AGENTS.md instructions for microfeed/microfeed, covering repository guidance, development workflow, instance management and cloudflare deployment, source architecture and theme repositories.
microfeed CLAUDE.md
Claude Code instructions for microfeed/microfeed, a project described as: an agentic cms self-hosted on cloudflare, for podcasts, blogs, photos, videos, documents, and curated urls.