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/archivebox/abx-dl/claude-mdgit clone --depth 1 https://github.com/ArchiveBox/abx-dlWrote 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/archivebox/abx-dl/claude-md)<a href="https://agentmods.dev/instructions/archivebox/abx-dl/claude-md"><img src="https://agentmods.dev/badge/instructions/archivebox/abx-dl/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 | $0.00547 | $0.00547 |
| Opus 5 | $0.00273 | $0.00273 |
| Sonnet 5 | $0.00109 | $0.00109 |
| Haiku 4.5 | $0.00055 | $0.00055 |
Grade A, and why
abx-dl CLAUDE.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
plugin_info="$(uv run --no-sync --no-sources abx-dl plugins wget)" How it starts
The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Instructions for abx-dl
Use AGENTS.md as the canonical repository guide. Keep this
checkout on main, use uv for Python commands, and exercise real CLI,
plugin, subprocess, network, and filesystem paths in tests. Do not add skips,
xfails, retries, fake hooks, fake binaries, fake buses, or fallback binary
resolution.
Development setup
set -Eeuo pipefail
uv sync --locked --dev --no-sources
help_output="$(uv run --no-sync --no-sources abx-dl --help)"
grep -q 'Usage:' <<<"$help_output"
Inspect the installed plugin surface
set -Eeuo pipefail
plugin_info="$(uv run --no-sync --no-sources abx-dl plugins wget)"
grep -q 'WGET_BINARY=wget' <<<"$plugin_info"
grep -q 'on_Snapshot__35_wget' <<<"$plugin_info"
Verification
The normal CI workflow assigns every repository test file exactly once across
Linux, macOS, and every supported Python minor version. It also runs every
non-Docker documentation snippet and a real crawl across every Snapshot hook
from the pinned abx-plugins release. The Docker workflow builds and tests both
amd64 and arm64 images and runs every snippet marked docker_required.
Run the repository's complete static verification before publishing changes:
set -Eeuo pipefail
uv run --no-sync --no-sources prek run --all-files
Releases
Releases are automatic from main. Bump the version and its lockfile, commit,
and push. The release workflow waits for the exact commit's test workflow,
creates the matching GitHub release and tag, waits for the exact Docker build,
and publishes the package through PyPI trusted publishing. Do not create or
push tags locally.
Architecture
- Plugin discovery comes from the installed
abx-pluginspackage or the explicitABX_PLUGINS_DIRoverride. - Required binaries are emitted as
BinaryRequestEventrecords and resolved only byabxpkg; host binaries are considered before managed providers. - Resolved host and managed binaries are projected into
ABXPKG_LIB_DIR/env/bin.ABXPKG_LIB_DIR/binis only a human convenience surface and must never be used programmatically. - Crawl setup and snapshot hooks run through the unified event state machine. Chrome readiness uses its plugin-owned file mechanism rather than a generic bus-level readiness event.
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.
- 5d ago First seen · 62 lines · 547 tokens per session scan A 22d3f89a0a7b
abx-dl CLAUDE.md is an instructions file published in the GitHub repository ArchiveBox/abx-dl (141 stars, last pushed yesterday), licensed MIT. It adds 547 tokens to every session, about $0.0027 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
meerkat AGENTS.md
AGENTS.md instructions for lukacf/meerkat: This repository uses Make as the developer-facing command surface.
appwrite AGENTS.md
AGENTS.md instructions for appwrite/appwrite, covering appwrite, commands, stack, layout and libraries.
wordpress-nginx-docker-compose CLAUDE.md
Claude Code instructions for urre/wordpress-nginx-docker-compose, covering claude.md, what this is, architecture, two .env files — don't confuse them and common commands.
workshop copilot-instructions.md
Copilot instructions for canonical/workshop, covering github copilot instructions, project overview, tech stack, repository structure and core directories.
metamask-extension AGENTS.md
AGENTS.md instructions for MetaMask/metamask-extension, covering agents.md, agent instructions summary, critical rules for agents, comprehensive guidelines location and quick setup.
android-qa-kit AGENTS.md
Instructions for willbytee-sudo/android-qa-kit, covering agents.md — setting up and driving an android device, 1. how to work with the user, 2. start by finding out what's already there, 3. what you can do in each mode and 4. if there's no device: ask first.