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/moq-dev/moq/claude-mdgit clone --depth 1 https://github.com/moq-dev/moqWhat 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 | $0.07735 | $0.07735 |
| Opus 5 | $0.03868 | $0.03868 |
| Sonnet 5 | $0.01547 | $0.01547 |
| Haiku 4.5 | $0.00773 | $0.00773 |
Grade A, and why
moq 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 today.
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 — 246 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance for AI coding agents when working with code in this repository.
Project Overview
MoQ (Media over QUIC) is a next-generation live media delivery protocol providing real-time latency at massive scale. It's a polyglot monorepo with Rust (server/native) and TypeScript/JavaScript (browser) implementations.
Common Development Commands
# Code quality and testing
nix develop --command just check # Lint and compile what the branch changed
nix develop --command just test # Test what the branch changed, same scope
nix develop --command just fix # Auto-fix lint/formatting, same scope
nix develop --command just check-all # Same as check, over every package
nix develop --command just fix-all # Same as fix, over every package
nix develop --command just build # Build all packages
nix develop --command just bench # Benchmark the current tree
nix develop --command just bench BASE # Compare BASE with the current tree
Use the Nix dev shell for project commands so local runs match CI tooling. If Nix is unavailable, use cargo or bun directly.
just bench runs every Criterion target plus fixed video and 1:N fanout
workloads against a temporary local relay. Pass a commit or ref to benchmark that
revision first and print base-to-current changes. Timing changes are informational;
benchmark crashes, zero delivery, and invalid samples still fail the command. Relay
CPU and RSS are included on Linux, where moq-bench-host can read /proc.
just check, just test, and just fix all diff the branch against its base and touch only the crates that changed plus everything depending on them, which is what keeps them fast when several worktrees are building at once. They skip a language entirely when the diff doesn't touch it. Reach for just check-all / just test all / just fix-all when you want the unscoped suite. They also switch to it themselves once the changed-file list outgrows what a single argv/env string can hold (64 KiB, roughly 1500 paths), because the list is passed to each per-language recipe as one argument. See Workflow for how the base is resolved.
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.
- today Changed · +11 lines · +368 tokens per session 1af581e78d2e
- 2d ago First seen · 235 lines · 7,367 tokens per session scan A a36efad36b65
moq CLAUDE.md is an instructions file published in the GitHub repository moq-dev/moq (1,486 stars, last pushed today), licensed Apache-2.0. It adds 7,735 tokens to every session, about $0.0387 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
streamkit AGENTS.md
Instructions for streamer45/streamkit, covering streamkit agent notes, what is streamkit, codebase map, tech stack and workflow.
streamkit CLAUDE.md
Instructions for streamer45/streamkit, a project described as: StreamKit is a self-hosted real-time media processing engine with pluggable node-based pipelines.
pipecat AGENTS.md
AGENTS.md instructions for pipecat-ai/pipecat, covering agents.md, project overview, common commands, setup development environment and install pre-commit hooks.
azure-sdk-for-java copilot-instructions.md
Copilot instructions for Azure/azure-sdk-for-java, covering prompt for github copilot, repository purpose, repository overview, prerequisites and azure sdk mcp server: quick start.
azure-sdk-for-java AGENTS.md
AGENTS.md instructions for Azure/azure-sdk-for-java, covering agents.md, repository purpose, repository structure, agent capabilities and supported actions.
next-sanity AGENTS.md
Instructions for sanity-io/next-sanity, covering agent notes, next-sanity/live export conditions, runtime nuance, ai literacy framework (ailf) and cursor cloud specific instructions.