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/sinanpl/mcp-memes/claude-mdgit clone --depth 1 https://github.com/sinanpl/mcp-memesWrote 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/sinanpl/mcp-memes/claude-md)<a href="https://agentmods.dev/instructions/sinanpl/mcp-memes/claude-md"><img src="https://agentmods.dev/badge/instructions/sinanpl/mcp-memes/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 | $0.01509 | $0.01509 |
| Opus 5 | $0.00754 | $0.00754 |
| Sonnet 5 | $0.00302 | $0.00302 |
| Haiku 4.5 | $0.00151 | $0.00151 |
Grade A, and why
mcp-memes 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 — 145 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for Claude Code working in this repository.
What this is
An MCP server that hands a model a catalogue of meme formats, lets it pick five and caption them, and opens a browser-side editor as the deliverable. Deployed as a Cloudflare Worker. See README.md for the architecture and docs/deployment.md for the deploy.
Build first, always
npm run build
scripts/build-widget.mjs bakes the editor, the Anton font and the server icon
into src/generated/assets.ts, and that file is gitignored. src/server.ts and
src/render.ts import it, so without a build nothing compiles and nothing runs —
not the tests, not the probe, not wrangler deploy. This is the single most
common way to be confused in this repo.
npm test # vitest, ~30 tests
npm run typecheck # tsc --noEmit
Invariants — do not break these
One copy of the layout rules. src/layout.ts is the only place layout logic
lives. It is bundled into the browser editor and imported by the Node renderer,
which is why what you edit is what you get. Never add layout maths to
widget/main.ts or src/render.ts; both must stay callers. src/layout.ts must
also stay pure — no DOM, no node: imports — because it runs in both. Text
measurement is injected for exactly that reason.
No matching algorithm. Choosing formats is the model's job, deliberately.
Do not add embedding search, scoring, or "best template" heuristics. The
catalogue is exposed as readable resources and the model picks. resolveTemplate
is identity mapping (an alias resolves to exactly one template or to nothing),
not fuzzy matching; suggestTemplates only produces error-message hints.
The core is host-agnostic. src/server.ts knows nothing about transports.
src/worker.ts, src/http.ts and src/stdio.ts are adapters over it. New host
support means a new adapter, never a change to the core. Only src/http.ts may
import express.
The editor is the answer. make_meme returns a deliberately content-free
line annotated audience: ["user"], priority: 0.1, and the picks travel in
structuredContent where the app reads them and the model does not. This is on
purpose: naming the formats in the text invites the model to list them back in
prose, which pushes the editor panel up out of the last turn. Do not "improve"
that message by making it informative.
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 · 145 lines · 1,509 tokens per session scan A fff5fdd1a267
mcp-memes CLAUDE.md is an instructions file published in the GitHub repository sinanpl/mcp-memes (0 stars, last pushed 6d ago), licensed MIT. It adds 1,509 tokens to every session, about $0.0075 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
mcp AGENTS.md
AGENTS.md instructions for cloudflare/mcp, covering agents.md, project overview, mcp specification compliance, repository structure and setup.
mcp-ts-core AGENTS.md
AGENTS.md instructions for cyanheads/mcp-ts-core, covering developer protocol, consumers, core rules, exports reference and import conventions.
mcp-tts-voicevox CLAUDE.md
Claude Code instructions for kajidog/mcp-tts-voicevox, covering claude.md, commands, build & validate (whole workspace), per-package and run a single test.
flaim AGENTS.md
AGENTS.md instructions for jdguggs10/flaim, covering repo agent guide, scope, public/private boundary, important distinction: product assets vs repo instructions and start here.
chapplin AGENTS.md
AGENTS.md instructions for ssssota/chapplin, covering agents.md, goal, project summary, session bootstrap and repo map.
mcp-migration-check AGENTS.md
AGENTS.md instructions for AlpayC/mcp-migration-check, covering working in this repository, shape, rules that are not negotiable, before you say it works and adding a rule.