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/techjewel/mdread/claude-mdgit clone --depth 1 https://github.com/techjewel/mdreadWrote 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/techjewel/mdread/claude-md)<a href="https://agentmods.dev/instructions/techjewel/mdread/claude-md"><img src="https://agentmods.dev/badge/instructions/techjewel/mdread/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 | $0.03109 | $0.03109 |
| Opus 5 | $0.01554 | $0.01554 |
| Sonnet 5 | $0.00622 | $0.00622 |
| Haiku 4.5 | $0.00311 | $0.00311 |
Grade A, and why
mdread 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 3d 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 — 176 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What this is
A local-first Markdown reader/editor. No framework, no tests. It's a Vite + SCSS app
of plain ES modules; files are read in the browser and nothing is uploaded unless the
user explicitly shares. vite build bundles src/ + public/ into dist/, and dist/
is what Cloudflare serves.
There is now one small backend under worker/, running only for /api/*. It is
deliberately blind in two places, and both are load-bearing:
- Shares — the browser seals a bundle with AES-GCM and the key travels in the URL fragment, which is never transmitted. Do not "improve" this by moving encryption server-side or by putting the key in a query string or path segment.
- Vault (synced share links and documents the user explicitly saved) — encrypted
with a key stretched from the user's master password via PBKDF2, client-side. The
server stores
SHA-256(email)and blobs. Do not add a password-reset flow: there is nothing to reset, and adding one would mean holding the key.
Reading and editing local files still never touch the network — the vault is opt-in per
document, via Add to vault in the topbar. That button lives with the document it acts
on, not in the sidebar; the sidebar panel previews the five most recent and hands off to
#vaultDlg (the full, searchable vault) via "Show all".
Vault storage is split deliberately. The index (vault:<acct>) holds share links
plus a document listing (id, name, path, size, updatedAt). Each document body is its
own key (vaultdoc:<acct>:<id>), so saving one document doesn't rewrite the whole blob
and the index stays cheap to fetch on unlock. Bodies are never cached in plaintext
locally — only the index is, so the sidebar can render before anything is decrypted.
Naming: the product/UI is "mdread" (and
wrangler.jsonc/package.jsonname), but the directory, code comments, IndexedDB database, andlocalStoragekeys all usemarkread. Both refer to the same thing — don't "fix" one to match the other.
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.
- 3d ago First seen · 176 lines · 3,109 tokens per session scan A ae58d398239d
mdread CLAUDE.md is an instructions file published in the GitHub repository techjewel/mdread (22 stars, last pushed 1mo ago), licensed MIT. It adds 3,109 tokens to every session, about $0.0155 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 2026-08-30.
Other instructions, from other repositories
Knote AGENTS.md
Instructions for 1661169091kiwi/Knote, covering agents.md, 构建与验证约定（用户明确要求，所有会话必须遵守） and 测试说明.
siyuan AGENTS.md
AGENTS.md instructions for siyuan-note/siyuan, covering agents.md, 1. non-negotiable constraints, do not hand-edit, verification and prohibited operations and 2. project-specific rules.
fieldtheory-cli CLAUDE.md
Instructions for afar1/fieldtheory-cli, covering claude.md, commands, architecture, key files and data flow.
syncular AGENTS.md
Instructions for syncular/syncular, covering syncular agent instructions, doctrine (enforced), coding rules, prose rules for docs and voice.
anytype-ts AGENTS.md
AGENTS.md instructions for anyproto/anytype-ts, covering agents.md, project structure, state management with mobx, electron integration and inter-process communication (ipc).
neurostack CLAUDE.md
Instructions for raphasouthall/neurostack, covering neurostack - claude code guide, quick reference, installation, mcp server (recommended for claude code) and openai-compatible api.