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/bassotov/deck/claude-mdgit clone --depth 1 https://github.com/bassotov/deckWrote 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/bassotov/deck/claude-md)<a href="https://agentmods.dev/instructions/bassotov/deck/claude-md"><img src="https://agentmods.dev/badge/instructions/bassotov/deck/claude-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 | $0.04896 | $0.04896 |
| Opus 5 | $0.02448 | $0.02448 |
| Sonnet 5 | $0.00979 | $0.00979 |
| Haiku 4.5 | $0.00490 | $0.00490 |
Grade A, and why
deck CLAUDE.md 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 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.
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 — 353 lines — stays where its author put it; the contents beside it link to each section on GitHub.
deck
Storylines are markdown files; this app renders them as 16:9 slides you can scroll
through and draw on while recording a voiceover. README.md is the user-facing doc —
this file is the working reference for the internals.
Run it
npm run dev # vite, http://localhost:5290 — the app-development loop
npm start # build, then the CLI, serving the built app
npm run mcp # build, then the MCP server on stdio
npm run typecheck # both projects
Shape
Three front ends over one core. The rule that makes this work: nothing product-shaped
lives in vite.config.ts, and nothing transport-shaped lives in server/store.ts.
server/paths.ts package root, dist and the export script, found not counted
server/config.ts where the decks are — argument, then env, then env file, then ./decks
server/slides.ts the slide grammar. imports no `node:`, so it bundles to the browser too
server/store.ts every filesystem operation. no HTTP, no MCP
server/api.ts the HTTP surface over a store; resolves true when it handled a request
server/serve.ts the standalone server: built app + api + live reload, no Vite
vite.config.ts the dev adapter — middleware around the same handler
bin/deck.ts the CLI
mcp/ the MCP server and the syntax crib it hands to agents
src/ the React app
scripts/export.ts PNG export, driving `serve` through Playwright
scripts/bundle.ts the Claude Desktop extension (.mcpb), staged and packed
mcpb/icon.html the extension icon's source; the png is rendered from it
The extension
npm run bundle writes build/deck-<version>.mcpb — a zip of manifest.json, lib/,
dist/, an icon, and a production-only node_modules installed into the staging
directory. It has to be self-contained: the host runs node lib/bin/deck.js mcp
directly, with no install step and no guarantee the user has ever run npm.
The decks folder arrives as DECKS_DIR from a user_config field of type directory.
An optional field the user leaves blank is passed through as the literal string
${user_config.asset_root}, so resolveConfig treats any value still containing ${
as unset — otherwise it would resolve to a directory with ${ in its name.
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 · 353 lines · 4,896 tokens per session scan A e274d5a60e9a
deck CLAUDE.md is an instructions file published in the GitHub repository bassotov/deck (2 stars, last pushed 1mo ago), licensed MIT. It adds 4,896 tokens to every session, about $0.0245 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 instructions, from other repositories
research-presentation-template CLAUDE.md
Instructions for adithya-s-k/research-presentation-template, covering claude.md — authoring this deck, the three things people get wrong, where things live, adding a slide and colour rules.
PPTAgent CLAUDE.md
Claude Code instructions for icip-cas/PPTAgent, a project described as: An Agentic Framework for Reflective PowerPoint Generation.
mindle AGENTS.md
AGENTS.md instructions for nonatofabio/mindle, covering agents.md, what mindle is — and isn't, tenets, architecture in 60 seconds and coding style.
exomem AGENTS.md
AGENTS.md instructions for Artexis10/exomem, a project described as: Self-hosted MCP server that makes your Obsidian/markdown vault searchable — text, PDFs, Office docs, images, audio — from any MCP client. Hybrid retrieval over a typed, governed corpus, sub-second at 50k notes; your files stay plain markdown.
agentic-resume-builder CLAUDE.md
Claude Code instructions for WtotdeD/agentic-resume-builder, covering agentic-resume-builder development guidelines, what this repo is, active technologies, package manager and validation.
dbv-md-reader CLAUDE.md
Claude Code instructions for davidbuenov/dbv-md-reader, covering instrucciones del proyecto para claude code — dbv-md-reader and ⚠️ reglas core.